Oracle Corporation (ORCL) Earnings Call Transcript & Summary

Please welcome Ken Bond to the stage.

Hello, and welcome to the Financial Analyst Meeting for 2024. Been a very exciting week. Attendance was definitely up. Energy, way up, way up. Now there's a lot of thematics you'll see through the show, one of which is artificial intelligence. Now using artificial intelligence, I posed the question, what slide would investors want to see most? And here's what it came up with. I know. I know. You love it. You love it. I know. Me, too. Okay. Kidding aside, I need to call out some important key points about our safe harbor, right? First, and obviously, we will be making some forward-looking statements today. And of course, the statements are subject to risks and uncertainties. Some of the factors that could affect our forward-looking statements, we detail those for you in our financial filings with the SEC, including Forms 10-K and 10-Q. All the information we're sharing with you today is as of today. Now lastly, we will not be undertaking any obligation to update the forward-looking statements we'll be making today in light of future information or any other market moving events, we will not be obligated to update what you hear today. Second thing is, we will be making use of non-GAAP financial measures. So just keep that in mind. They'll be noted for you on the slides when we're using non-GAAP measures. And then, of course, especially with some of the product guys, you will hear some comments about some of our technologies. But just bear in mind, these are just for informational purposes, there's no commitments about any of this. Just keep that in mind. Okay. The agenda. We're going to have a pretty good day. Busy. We'll start. Clay is going to come on out first, talking about Oracle Cloud Infrastructure. And then after that, we're going to have a couple of panels, actually 3. Two roundtables led by Leah Yomtovian. She'll be basically bringing up the first group talking with you about database and analytics. Following that, we'll bring up our apps guys, including Mike Sicilia, running the industries business, and they'll talk about cloud applications. And then it's always good to hear the voice of the customer. Safra did a great job just the other day, bringing up tons of customers, better that you hear it from them sometimes. And we'll have some of those folks coming up from Nomura Research Institute, MGM Resorts and Vodafone. That will take us to lunch. After lunch, Doug will come on up and provide you all with a financial update. And following that, Safra will come on up. She's excited about doing the Q&A. She wants to hear all the great questions you have. And then we'll do our usual annual Q&A with Larry, who will definitely be here. So with that, let's go ahead and get started.

Please welcome, Clay Magouyrk, to the stage.

That was less of a reception than I was expecting. Can we -- I'm going to come -- I'm going to walk over here again, and I'm going to need -- okay. See, that's not even -- that's not for me, that's for you. Look, let's be honest. I'm going to get to be done here in about 30 minutes. You're going to have to sit here and the higher energy you can maintain, I don't know if any of you have a WHOOP on, your calories burned will be higher. Okay. So this is the slide that Ken mentioned that he says I have to also tell you. And then I think this is the second slide that Ken mentioned that he says, I also must tell you about. So those are the slides that Ken talked about. OCI is growing very quickly. Based on which key metric you want to look at, if you think about revenue growth up more than 50% year-over-year in terms of customer-facing regions, and the reason customer-facing regions is important, just so your understanding is we don't build regions, unless people want them. I've tried asking Safra, if she'll let me just build regions that don't have demand. You don't want to hear her response. But the fact that we are growing the region footprint this quickly is a very different way of looking at our business. The fact that data center capacity is growing at a much faster rate than revenue. We're not building data centers because we don't have demand either. So these are all indicators of the scale of the demand that we're seeing across the industry. Another key metric that I think is important to grasp is that the number of customers that spend more than $5 million a year with us on OCI directly. This doesn't include the other parts of our business. It's not talking about database, license revenue support or -- but just on cloud infrastructure has grown 42% year-over-year. So what is it that's driving this growth? And for those of you who've been here before, which is many of you, I really try to take our business and break it down into a few different pieces. So we're going to talk about our enterprise customers, which are the ones that everyone knows the most about here at Oracle, we've had for a long time. We'll talk about cloud native and AI customers. And then we'll talk about another critical part of our business, which is what we're doing with distributed cloud around dedicated region and Alloy. So when you look at what's driving enterprise growth, we can talk about many of the services, and I'll bring some of those up in a second. The general things that you'd expect from any public cloud provider. You launched more regions, you launch more features and functions, you scale your sales organization, you scale your ability through SIs. We're doing all of that. But the single biggest thing that we've done recently is that if you go back 13 months ago, there was one cloud where you could get Oracle Database services. There was one cloud that had its salespeople out selling Oracle Database services. And it was a great cloud. Look, I'm very proud of it, but it was only one cloud. As of today, we now have 4 clouds. And it's not just any 4, it's the 4. And I don't know how many of you have talked to different customers. I've had a lot of different conversations. There's something very different about some versus all. And when I have conversations with customers now, they are incredibly encouraged by the fact that it doesn't matter which cloud they choose. They can maintain their investment in the Oracle database, they can move that into the cloud and still get all of our best and greatest services. This, along with all the other investments we're making into our general purpose public cloud, is what's causing us to have such high growth, for example, and the number of customers spending more than $10 million in the enterprise space year-over. And I want people to also grasp -- I'll talk about the number of regions we're launching here in a minute. We have several thousand racks of capacity. And by capacity here, I don't mean we land the racks ahead of time. What I mean is that we have plans for -- this is the size of the market that we see in our conversations with ourselves and with our partners. If you take a look at this map, you can see a few things. One is that we need a bigger screen because there's not enough place to put all the dots. But we have a lot that we've been doing both across our commercial regions, our dedicated cloud regions. But when you add to it, the huge investment we're making across our multi-cloud partnerships, you realize that the available locations in which you can procure our industry-leading, best-in-class, highly differentiated data platform services, you're just able to procure them in so many more places than you were yesterday. I'm going to take a moment and try to get everyone here to understand one critical point. This is something that if you get it, you're going to feel like I'm spending 3 minutes on a topic that is obvious. But I've talked to enough customers now who, even after I've said this, probably is a feedback from me, I put it in my performance review, get better at explaining technology to people. But it's really important to understand the way in which our multi-cloud strategy is different than what anyone has done before. So if you go and you look at a traditional company, I'm not going to name names, but if you think about companies that are doing analytics on Amazon and Microsoft and Google or companies that are doing databases, the way in which they run is they have a service and they run that thing on top of those clouds. So you go to Amazon and you procure some EC2 instances and you procure some elastic block storage and you procure some S3 storage and you build your application, your service. And you -- and then when you want to do Google, you have to go and redo that thing, and you have to use it on top of that cloud. And the same thing applies for Microsoft Azure. That's not what we've done with OCI. What we've done is we've created a first-in-class kind of ever done before partnership where we bring OCI and we extend it into the other cloud provider. Okay. Now the reason that's important is that it's important for us, it's important for customers, and it's actually really important for our partners, and I'll explain why. First, the most important reason why this is so valuable to customers is that, I don't know if anyone has paid attention to the things that Oracle has been doing over the past 47 years. And I know that in a few minutes, Leah is going to be on stage, and we're going to hear from Edward and Juan and TK about all of the great innovation. Everything that we have available inside OCI is now going to be available in our partner clouds. So as we come up with a new hardware platform, the next generation of Exadata, the next generation of our highly optimized network, I don't have to go and try to bargain with these multi-cloud partners to allow me to use that technology. It's my technology. I put it in there, okay? The reason customers love that is that they get the same level of service and quality. It's not a matter of, well, it's great. We rolled it out to OCI, but it's going to take us 18 months to get it into Google and then -- no, it's the same service. So when I was having a conversation with customers yesterday and they said, hey, Clay, so like what are your references for these recently launched 4 regions with Google? I go, Fusion, NetSuite, Exelon, FedEx, every single customer that runs their databases in our cloud today. And they go, I don't understand, it's the same service. I don't mean it's like -- it's not even a separate copy of the same service. It's literally the same service with the same functionality. Now I know -- again, it feels like there's a horse and it was -- it was no longer alive, and I was hitting it. I understand. But it's important everyone here understand the true value of that, the customers get all the best capabilities that we have instantly. We love it because it massively simplifies our management. We don't have to go in and have 4 copies of this service with extreme cost and overhead, it's the same thing. And our partners like it because what they know is that they actually get the same quality across all of them. I don't have this conversation with Microsoft or Google about when do I get to say, great, it's the same. They're happy because they get the same capabilities. So I'll stop beating that one. But it's really important everyone walk away understanding why that's so different. We also continue to launch more and more new services. our Cashing service, something -- I'm not sure if Juan will talk about the fact that we now have our globally distributed autonomous database, I think, is truly incredible. We continue to innovate across the board and all of these services end up right -- continuing to grow across our enterprise customer base. So moving on to our cloud native and AI customers. 162% year-over-year growth in the number of customers in this category, more than 2.5x capacity increase compared to where we were a year ago, serving these customers and more than $3 billion in TCV of wins in Q1. Why do these people pick us? Well, it was funny. I'll tell you an anecdote. As part of the preparation for this, I was talking with Doug Kehring, who talks later, and he sent me a text. He's like, hey, what was the exact date when we launched OCI. And so I type it into Google and I found the blog post that was written in 2016. And I've been working on OCI from the very first day until now. And it was interesting to go back and to read the blog post. And I was like, wow, either we were -- either we're not good at changing our strategy or the strategy was good because I can tell you to a blog post from 8 years ago that talks about the same things that I'm talking about today. Now it's much easier to convince people it's a good strategy that we're successful. I can tell you is that when you're talking about a strategy and you haven't been successful yet, it doesn't always go well. These customers are choosing us because of very fundamental things like performance and availability, security and support. That's what it is. And if you go back and you look at our very first launch, and we said, here's our first region, that's what we talk about, okay? A couple of notable ones here. Over the past quarter, cloud flare -- sorry, CrowdStrike and Palo Alto joined us. The reason for that is because they're becoming more and more aware of the economic advantages and the security advantages that our cloud offers. And then if you were seeing -- saw my keynote, you can see how a company like Skydance has now rendering entire films on top of us. But then we're working very closely with Greg and the team on actually completely revolutionizing their entire studio where they don't want to have any more workstations. They don't want to have any more of their own data centers, everything to be able to run an entire animation studio runs in the cloud. That's the breadth and depth of this segment of our customer base. So it's great that we've been talking about it for a long time, but why? Why does it work? I'm not going to stand up here and talk about a whole bunch of very detailed technical pieces. I will talk about one that hopefully will resonate. So in the very beginning, we decided to do BareMetal. We offered -- in the beginning, actually, we only offered BareMetal. And we did that because we believed it was the right security posture, such that if you have that clear separation between what you can give a customer and you can firmly believe you can take it back and securely wipe it, you build a more secure cloud over time. We also, in the beginning, invested very heavily because of the great work that Juan and Andy and their team did on Exadata, to ensure that we've built dedicated RDMA networks for clusters, right? If you were at any other company and you were building a cloud from scratch, I promise you that would not have been a priority for you, okay? And by the way, when we did that technology, we didn't just build it. We made sure it just as secure, that it was entirely virtualizable in a different way than our front-end network, but it was multi-tenant so we can carve it up and we can -- so we did all of that. So suddenly, along comes AI. And what does AI need? AI needs clusters. They might not be bigger, they might be smaller, they might look different, but they need highly secure, very performant clustering technology. If you actually go and look at the available clouds today, ourselves and our key competitors, and you actually go and understand, why are people choosing us? It's because of these same investments. It's the fact that we've built a completely virtualizable RDMA network, but it's integrated into our systems. So what other cloud providers will do is they'll say, oh, we have this, but it's basically BareMetal hosting off to the side. That's not what we do. We have complete API control, you can provision, you can restart, you can securely wipe. Other people don't do that. Many other clouds don't even have a dedicated network. They said, hey, that's too complicated. It's too expensive. It's too hard. We don't want to deal with all that complexity. Why don't we just sacrifice your performance and run it over the same front-end network, right? So when this workload came around, yes, we had work to do. I'm not saying that we didn't have to design larger scale networks, and there isn't different tuning [ objects ] but we can go into all of ECNs and how you do congestion control in a very large [ Gfab ] and how does that compare to a couple of Exadata racks. But what I can tell you is the fact that at Oracle, we believe that performance matters and we believe that security matters and that you don't just tack it on at the end, that you make this a critical part of your functionality, enabled us when this opportunity came around, right, to provide just a better offering to our customers. And you see this in our numbers every day. I mean, I -- this is -- I update this slide just so I don't copy pace the same slide year-to-year. I do actually have my team go out and make sure because prices can change. You can come to OCI and at list price. You don't have to talk to me. You can do it with a credit card. We have many -- Oracle, we're really good at taking different forms of payment. I don't know if you knew that. But if -- we try to never turn money down as long as it's accurate and legal way to do it, obviously. Or you can go to our competitors, and you can commit all upfront for 3 years and you're still at least 10% more expensive than if you just came to OCI. This is why these customers pick our cloud. And most of my time is not spent modifying our technology to make it work for them. Most of my time is spent them going, I don't believe you. And I go, okay, well, but you might want to try it. And they go, well, still -- okay, and they try it and they go, oh, this is actually really, really good. I was like, yes, it's almost like that's the thing we've been saying, remember, for the whole time we've been here. Another area that we're investing very heavily is that -- so I talked a bit about the networking side and why we're highly differentiated there. The fact that we can move very quickly, the fact that we can offer the latest and greatest of all of NVIDIA's GPUs, the fact that we're investing very heavily into our managed luster offering, the fact that we update our file storage service to be able to be used for scale-down customers, the fact that we offer AMD GPUs. And we have a huge amount of our energy dedicated to enabling these types of customers, right? And who's who of logos on this slide, and there's many more that I can't show you publicly, right? These are the people pushing forward the frontier of what's going on in AI, right? Those people are -- they're not uninformed technologists. They know their business and I promise you, they're not coming to us because they haven't heard of our competitors or they haven't heard of any other options. They try what we have, and they're extremely impressed. So hopefully, that segment of our business, it's clear why that's continuing to grow at an accelerating rate. So I want to take a few minutes now to talk about distributed cloud and Alloy. So one of the other things that we built into OCI from the very beginning was a different belief about the way you should build clouds, the way you should build your regions, the way in which they should be operated and the way in which customers should be able to consume the cloud. Having worked at multiple cloud companies, I can promise you, and to be fair, it's nice having that late mover advantage sometime, it was -- none of the other cloud providers designed in from the beginning, the concept that someone might be able to run the cloud in their own data center. The idea was very simple. We'll build a relatively few in the order of 10, 20, 30 really large places, and all of the world's computing will happen in those 30 large places. We thought that was a bad idea. We thought that, well, actually, a, it's a concentration risk, you shouldn't put all of your computers in one space. We thought that there's this thing called countries, and they care about where their data lives. We found these things called customers, and they care a lot about being able to bring the cloud to them. And so we've been on a journey for the entire time that we've been around, focused on how do we scale down as well as scale up, how do we offer customers choice about deployment. So we started with dedicated region, which was -- it's not a genius idea. It was, well, if we just make the cloud small enough, there are customers that would like one. And so we worked very hard to scale it down, and that required a bunch of work on the way in which we build and operate these clouds. I cannot tell you the amount of effort that we put in to be able to push a button, and then at the end of that button push, you have a cloud region. I know it's -- you're going to think, well, Clay, isn't that obvious? That's how you would do it. No. The way that most of these cloud providers build the cloud is it's the same way that when a city gets an Olympic win, they build the Olympics. They go in and they go, great, we have a construction project and we're going to build this thing, and there's project managers and there's people manually employing staff. And at the end of it, you end up with a one-off cloud. That's the way our competitors build their cloud. That's not how we build our cloud. We have an installer that runs and it runs. And at the end of it, it goes, oh, we brought up the cloud, it works. That's the only way we can build this very large number of clouds. But it's really hard because this is not a packaged software product. You're not building a copy of windows, right? You're not building Firefox or Chrome. You're actually trying to build this massively sprawling set of functionality that spans across this very disparate hardware systems. So we scaled it down and we worked on our software. But we continue to innovate in that area, right? If you saw me on stage yesterday, there were 3 racks. In some ways, 3 racks seems interesting. In some ways, not. The racks themselves is not the interesting part. It's the fact that there's only 3 of them. This is not launched yet. It will be available very soon, right, sometime next year. But we already have this version that runs in 12 racks today, and we're already bringing out the 3 rack version. We had to read this on the network. We had to converge our hardware platform. But this enables a huge opportunity for us because customers now suddenly can get the cloud. They can put it in their data center and they don't have to sit here and figure out, well, how do I do this complicated move of everything over there. They can just bring the full cloud functionality to themselves. And then we get other really interesting things. If you bring the security down to the rack level, such that now you can put the racks wherever you want, the cables, the way in which they're brought in, everything is encrypted when it leaves. And you actually -- from the outside, you can't get in and plug things into the computer. And we have real-time telemetry on the racks themselves. It's -- suddenly, it's very easy to take this cloud product, put it into your data center, wire it up and boot the thing up. The same thing we've been working on across what we call Oracle Alloy. Alloy is our offering that allows customers to become cloud operators themselves, right? If you were -- if you attended my keynote yesterday, you would see, I had Koga-san from Fujitsu on stage, talking about how excited he is that he now can offer all of his customers, not just in Japan, but globally, a cloud that gives them the control and data sovereignty that they need, but with the full features functions, right? In his conversation with me, he said, I had 2 choices before that were bad. I can run it on my cloud, which was great. I had lots of control, very secure, but did not have the functionality. I can run it in other hyperscalers, but I could not solve for the unique regulatory and control needs. Alloy solves that problem. And this is only possible because we have a dedicated region. If I go to Koga-san and say, hey, Koga-san, I'm going to need you to buy like 1,000 racks to start. He's like, man, this is a very hard business case to make. It's expensive, and I got to buy a lot of data centers. It's hard just to make it small. Suddenly, it's very easy to say, yes, let's get into that business. And it's not just Japan, right? You look at what we're doing with STC in Saudi Arabia, what we're doing with Team IM cloud in New Zealand. This is something that we have massive demand for around the world. And the other thing that I think -- I don't want to crib too much of what Larry always says, I'll repeat it though anyway. It's the fact that Oracle is both infrastructure and applications because to be able to offer Alloy as a complete solution, it's not just cloud infrastructure. I'd like to tell you that we've built all the important software in the world. It's not. We have a very busy team, but we're not nearly that good. If you want to be able to offer Alloy, you need an ERP system. You need CPQ. You need a way to bill your customers. You need to service cloud to be able to take the support request. Oh, here's the cool part. Oracle has that. So when we go with Alloy, we don't just sell cloud infrastructure. Included with that is fusion and other pieces of our business, our other industry applications that move together such that now a customer like NRI can use that technology to solve their customers' problems. The other thing -- this is the last slide, and then I'll get off the stage and we can go to Leah and the team. I've talked about these pieces of our business as if they're independent, and it's useful to talk about it that way. But there's also a lot of virtual cycle between them. So as an example, the types of customers that are procuring our dedicated region are also our enterprise customers. And the fact that we have our enterprise customer base is what allows, for example, our distributed cloud operators like Alloy that they then can move to that cloud. The same thing is true across our cloud native and AI customers. The fact that we've invested so very heavily into things like Exadata and high-performance computing for enterprise customers is what enabled us to actually be extremely good at operating the cloud for them. And what's also interesting is that as these cloud native and AI customers show up, they start off at times buying a lot of compute and storage and working, but they move up the stack. They start using autonomous database. They say, oh, I really would like to have fusion. And all of these things come together. So yes, we have these individual segments of our business, and it's important because they're -- they grow at different rates and their needs are different. But between them, there's a magical virtuous cycle where as we have more cloud native and AI customers, we get more scale, it lets us offer lower prices to our enterprise customers. The enterprise customers drive very complex demands. Our cloud business as a whole is very well diversified and growing in multiple different dimensions. And so with that, I thank you very much. It's been fun being here and hand it off to the next people. Thank you.

Please welcome Leah Yomtovian, Edward Screven, Juan Loaiza, and T.K. Anand to the stage.

Hi, everyone. It's great to be here. Before we get started, I think we have 2 slides from Ken that he shared with folks in this room before. Can we show those slides, just a reminder? Helpful reminder. Okay. Great. So we just heard from Clay on our business momentum in OCI and our key differentiators that are helping to accelerate our business. Now I want to turn to database and analytics. I want to pick up on the threads from Clay. Edward, I'll start with you on our platform services that we've built on OCI. We've taken a different approach as compared to others. Others have approached their services by building on all clouds. Our approach has been an integrated strategy where we build on OCI and deliver via multi-cloud. Why have we taken that different approach? And how is it helping our customers?

Well, first of all, it's very important that our platform services like database be available to our customers no matter which clouds they're choosing for the rest of their application stack. Now as you point out, I mean, one way to do that would be, we could just try to take Oracle Database and HeatWave and port it to every other cloud and try to make it run well there. But the problem is we would try to make it run well there. Because we're building on Oracle Cloud Infrastructure in our own cloud and the cloud regions that we attach to the other clouds, we can take advantage of a combined engineering between that physical infrastructure and the virtualization software that's running on top of it and the database software on top of it, whether it's Oracle database or HeatWave. So at the end, our customers get a service, which is a faster and far more functional and far more secure.

Great. And we offer flexibility across deployment options. We have 6 deployment options. It really enables us to be everywhere. Juan, I have a question for you, which is around our equity. How is it not only enabling us to maintain our database customers, but also accelerating the adoption of our unique capabilities like autonomous?

Yes. So that's pretty straightforward. One of the main reasons our customers come to our cloud is because we're running Oracle Database, the full Oracle Database, including our Exadata platform, which is what 90% of the largest companies in the world are already running on-prem and what they rely on. So having that available in all these clouds actually -- obviously makes it possible for our customers to adopt the cloud platform. So yes, we also have it on dedicated regions. We can provide it in cloud customer. We can provide them an Exadata on-prem. So yes, it's all there. It's back to our roots and database. We're everywhere. So that was one of the original roots of Oracle. It's available everywhere, everybody wants it. We're back to that. Very unique. This is a huge year for us in this respect. So yes, so that's -- it's very straightforward. And the other thing that Clay mentioned is true, which is this is the premier platform that all the big financial institutions, telecoms, retailers, everything. Sorry, my -- I've been talking all week and my voice is starting to run out. But this is it, and we're putting that exact platform, everything. And everything we built in the cloud for the last 8 or so years, autonomous database. We put a huge engineering effort into building that. It was exclusively in OCI. Now it's in all the major clouds. So it's a huge deal.

So we're meeting our customers wherever they are. We're making it easier for them to move to the cloud and take advantage of the unique capabilities that the cloud offers. I want to switch over to 23ai, which your team has been building, Juan. Can you talk about the value that we're delivering with 23ai? And why are we so confident that our customers are going to upgrade to 23ai?

Yes. So that's another very good point. So this is a very big year for database. And we talked about the first thing, which is we're everywhere now. We're in all the clouds. That's a huge deal. The second huge deal is in May, we released our latest major version of Oracle database that's called Database 23ai. So that's something where we've been working for several -- a lot of major architectural innovations in that release. Some of the stuff we've been working on for like 6 or 7 years, it's been kind of baking, baking, baking and it's finally come out. And there's 3 main focuses there. Number one, AI, AI. AI is huge. I don't have to explain that to everyone. AI is huge. Second was we put a huge effort into developers. A lot of innovative new development technology. And then our traditional mission critical, a lot of new developments in mission-critical technology. So we baked AI, things like vector search, natural language query, natural language search into the Oracle Database. We've adapted the Oracle Database to make it friendly toward AI. So there's AI in the database, but we've also changed the structure of Oracle so that generative AI, basically, it's easier for it to generate code for Oracle Database. So those are both very big things. In developers, we've unified some of the major data models that developers have been using, the document adjacent relational model, the graft model and the relational model. No one's ever done this before. This is a big breakthrough in data management that no one else has. So that was a major effort that we've had going on for several years. And there's also many, many enhancements on the mission-critical side. A big one we talked about Exadata. Exadata has really been for large customers, large enterprises. We have a new a technology called Exascale, which is kind of a software reengineering of the way Exadata work. And it scales even higher, but more importantly, it scales down. So now even a tiny little customer, the minimum size now for running Exadata is 2 cores. So even the tiniest customers now can get the benefit of Exadata. The thing I'd say is you can get stock exchange level performance, availability, security, even if you're the smallest customer. So again, that's very unique. We're bringing our super enterprise technology, down scaling at the midrange, even low-end customers. We have our global distributed database. Clay mentioned that. It allows customers to distribute their data anywhere they want around the world. This is huge for data sovereignty, more and more countries are passing laws saying data has to be local. I can go on, but I'm going to stop, right there. So much technology. So this is a huge year for those 2 reasons for database.

And why are we so confident our customers are going to make the upgrade and that it's going continue to accelerate our business momentum?

Yes. So because these technologies developed are really powerful technologies. And the one thing I would say is the AI technologies were helped by the fact that there's so much AI stuff going on in the market. But we -- our pre-release program for the AI technology is the biggest that we've ever had. We have more interest than that. And the reason people like it is they can say, hey, I need to adopt AI. I know I got to do this to stay competitive. And what we're allowing them to do is to say, hey, you can take these AI technologies and just put them in your existing mission-critical databases. You don't have to change anything. It's just a new kind of query, a new kind of data type. And every feature, all the availability, security, scalability, all that stuff, because it's just a feature of Oracle Database, everything that we've built for the last 40 years, it works with the AI technology. So it's instantly mission-critical, so you can deploy it for anything. And that is really a big deal for customers. They want to go quickly. They have directives from their CEO, from their Board, we got to do AI. And so how do you do AI for mission-critical systems? Well, if drop it into the Oracle database, that's the fastest, easiest, safest path to do AI.

This is great. So we're meeting our customers wherever they are. We're enabling them to move to cloud faster. We're enabling them to adopt AI more quickly. Edward, I want to come back to you and ask you, the Oracle Database is just one of the services that we offer. We also have MySQL HeatWave and other PaaS services. Can you talk about the unique value that those offer? And how does that translate into our business momentum accelerating?

Yes. I mean, of course, a lot of -- especially a lot of our cloud native customers, they've chosen to use MySQL as their transaction store. And my MySQL HeatWave is combined MySQL base online transaction processing with extremely high performance, highly scalable in-memory analytics. And in a lot of ways, if you look at what we do in Oracle Database, look at what we do in HeatWave, we sort of have a build it in philosophy, right? I mean, we have an amazing range of capabilities that are part of 23ai, right? And in the same way in HeatWave we took about base, highly scalable in-memory analytics and we put on top of it, something we call auto ML, so automatic machine learning. We let customers who basically they don't have to know anything really about data science in order to build progression models, classification systems, recommender systems, anomaly detection. We added on top of that something called Lakehouse. So we can perform these extremely high-performance in-memory analytic queries, right, just pulling from files that are stored in ObjectStore, right? And then we also just released Gen AI capabilities. So now our customers can do very high performance vector database creation, right, integrated with the normal analytic queries. So I could do joins across structured data and similar researches that are exact, right? I can combine that with results from auto ML for doing something like anomaly detection. And then I can feed all of that into large language models for generation. And just to make sure that our customers have it all built in, we actually have an LLM that's built right into HeatWave. We just try to make it as simple as possible for our customers to do extremely sophisticated data analytics, right, at extremely high speed across a wide range of scale. So everything from a single node all the way up to 512 nodes.

And before I turn to T.K. to dive more into data and analytics, I just want to ask Juan, Edward, is there anything else that think we should highlight for the audience here today that's helping us to continue to deepen and expand our relationships, especially with our enterprise customers?

Well, look, I think enterprise customers like performance and low cost like everyone else. But one thing I think that's especially critical is security, right? Security that you can achieve in a simple way, right? So we've always been very, very focused at Oracle, I mean, long before cloud on making sure that we had powerful security features built into our products. When we built our cloud, the #1 priority was keep the data secure, right? And we've built, I think, a comprehensive framework of security features and security enforcement mechanisms that help our customers stay secure. Now if you listen to Larry's keynote, he talked about something called zipper. Zipper is the next stage of world-class security. So imagine being able to write relatively simple expressions, declarative rules about what -- who has to have access to what, who may have access to what. And the fundamental infrastructure of our cloud network enforces that security. It simply will not permit access to data unless you're supposed to get it. And that zipper security screen is -- it's enforced at the network level that's understood by Oracle Database, HeatWave and other data stores. And I think that is very appealing to enterprise customers. I don't think there's a sophisticated enterprise customer out there who believes that they can do a good job of securing their own infrastructure. They just don't. They know they can't. And the only parties you can do it are cloud vendors and the cloud vendor who can do it best is definitely us.

Thank you. Juan, anything before I turn to T.K.?

I can go on, but let me just add to the zipper thing. Zipper is very cool technology, very unique. The thing I really like about it is it puts the security all the way through the stack, in the network, in the database, all the way across the network. So you can say, hey, this is a, let's say, a support analyst or something. And that knowledge goes all the way through the stack, through the application tier. So it knows your support on us, what can you do, what can't you do. It goes into the database. It says, hey, you don't have access to the credit card information because you're a support analyst. So all the way through the stack, which is something that's been missing because we've had these isolated levels of security without the integration across all the stack. So this is really kind of breakthrough technology, and it's going to be very, very interesting to customer security comes up all the time. I mean, you see what happens when there's a breach in any enterprise. It's horrific kind of situation where -- I mean, I've gotten, I think, 5 letters myself about data breaches and, oh, yes, we lost all your data. So this is not good.

Absolutely not good. I want to turn now to data and analytics. Obviously, our customers are looking to maximize the power of their data, T.K. So can you explain how is Oracle uniquely positioned to help our customers unlock the power of their data but really maximize the potential of it?

Yes. I mean obviously, Oracle is a custodian of so much of our customers' valuable data, data in our databases, MySQL databases. Oracle applications in the cloud, on premises. So it's just natural that they come, look to Oracle to help them get value out of all of this data. And in OCI, I think we have an amazing comprehensive suite of data intelligence services that help them extract value from the data. First and foremost, with autonomous data warehouse, which is one of the flavors of Autonomous Database and HeatWave. We have 2 of the industry-leading analytical database engines. These are the engines that can help really analyze and go through all of your data and extract insights and all that. So we have a great foundation, unique differentiator with that, right? And of course, now it's available in other clouds as well. Many customers want to also leverage open source technologies like SPARC to be able do data science and other workloads. So we have an intelligent data lake service that we're offering in OCI that helps bring those capabilities as well. And Oracle Analytics Cloud, which is our visual interactive analytics offering, is another capability that has been growing rapidly over the past few years. Recently, at the Gartner Data Analytics Summit, we had this thing called the BI bake-off, where we had a large audience and the 3 vendors, we were 1 of the 3 vendors alongside Microsoft and Tableau up on stage. And there was -- we had like a bake-off with among the 3 of us and the audience rated us the best. And so we had -- that's another piece of the strategy. And then, of course, we have a growing suite of AI and data science services and OCI. So we're integrating all of these together into a more cohesive experience for our customers. The world of data analytics is sort of converged. Customers want a simple integrated experience across data lakes and data warehouses and visual analytics and AI, and that's kind of what we're doing in OCI. But the nice thing about analytics that -- the approach we've taken is it's also a multi-cloud approach. We recognize that our customers also have data in other clouds. So we can go reach data wherever it is. So for example, our Oracle Analytics offerings can go -- allow customers to go visualize and explore data in an Oracle Database also relate that with data, let's say, in Google BigQuery or Snowflake. So we're open also in that nature. Yes.

So the concept of being open extends also to data and analytics?

Absolutely.

That's great. So now I want to talk -- another unique feature of Oracle is our breadth and our depth, and the fact that we have an integrated product portfolio. So T.K., can you explain how you're integrating our analytics directly into the applications our customers use day-to-day, whether it's for the front office activities, back office industry, can you talk about that a little bit?

Yes. I mean, this is what actually gets me super excited because this is the unique differentiator that Oracle has that no other vendor has. I mean, you all know, and Larry often talks about the thing that makes Oracle Cloud unique, is we have amazing cloud infrastructure and databases and so forth. And we have amazing suite of applications that leverage this infrastructure and makes the infrastructure better. Well, that same principle applies to analytics and data intelligence as well. So we have a rich suite of horizontal applications with Fusion and NetSuite. We've got industry vertical applications like health care, life sciences, financial services, et cetera. And we -- Oracle, have deep understanding of all of the data within these applications. They're not just bits and bites to us. They're actually entities like customers, products, patients. And we understand what the data means. We also understand what sorts of insights that customers are looking to get out of this data. We also know what are the types of intelligent decisions and actions that they want to take based on whether there's an adverse event that they want to react to and so forth. So what we've done is we've built a suite of data intelligence applications. We have something called Fusion Data Intelligence for our Fusion apps. We have something similar for NetSuite. We have health data intelligence for our Oracle Health customers. We're also working on other industries. We -- at CloudWorld a couple of days back, we announced the energy and water data intelligence application for our utilities customers and we're going to keep going because as a data geek like, I just love the fact that we have access to all this amazing data across horizontal and vertical domains. And then we're offering SaaS analytics. The data -- we take all of that data, we massage it, we'd get it ready for analytics. We give customers prebuilt analytics infused with AI. And best of all, we connect the insights back into the applications. Ultimately, that's what customers want. Customers are like, yes, I can use all sorts of analytics and business intelligence tools to go look at data and analyze and explore it. And then it's left as an IQ test to the user to go figure out, okay, how do I go improve my business? Why don't you just help me do that? That's kind of what we're doing.

Actually, at Oracle, we're obviously a customer of your technologies. One of the ways we use it is to help us ensure that we can drive the productivity of our sales reps. Can you explain some of the ways our customers are using the analytics day to day?

Yes. I mean, like there are -- I'll give you an example in the case of health care, right? Like -- so our health data intelligence offering is being used by health care customers that are a mix of Cerner as well as even Epic customers, right? They really appreciate the fact that we can bring all of -- I mean, health care is just a space that I find, which is just so ripe for disruption through analytics. We're able to bring data from all of the different hospitals, health care providers, payers and so forth. And we're able to bring a unified notion of a patient's health history. And we use that to offer health care providers to kind of manage their patient population as a whole, understand what are the sort of trends that are happening. More and more in the health care industry, they're moving to a value-based care model. So providers are being asked to measure themselves in terms of how they're improving the health care of the population as a whole as opposed to just getting paid for services. So that's something that we're helping, and also possible because we have deep intelligence about all of this happening in the system. We also do things like point-of-care solutions, like a patients walking right up to a doctor for an appointment, and we can give all this intelligence to a doctor about what is the history, the recent history with that patient. What are some preventive actions that can be taken. This is just one example. We have similar solutions for finance, HR, all sorts of other domains. And again, because we're the custodians of business data for our customers, we can provide these deep solutions.

Right. And the key theme across all the use cases is that you're enabling whoever it is to be more productive, be more efficient and drive the best outcomes.

Absolutely. We're not just giving our customers a data platform saying, you can go put all your data and they run really superfast queries and then it's up to you what you do with the data. We're actually turning it into a complete finished solution. And that's something that only Oracle can do because we have the platform, thanks to our amazing database technology, but we also have our applications and the domain knowledge.

Very unique. Juan, I want to come back to you because we've talked a lot about how we're expanding our reach and deepening our relationships with enterprise customers, but you also mentioned developers. Can you expand on that a little bit more? How are we continuing to expand our reach in the development community?

Yes. So that's one of our main focuses for our next release of the current release, actually. I keep thinking of the next release. We released it back in May. It's production. So our Oracle Database '23, I mentioned, I think the biggest thing is this model change is -- model changes don't happen very often in data management. This whole unifying the document graph and relational model is a huge deal. Another really big deal is our APEX development tool. We've been working on that for 20 years, and we're starting to build all our major apps on that. So this is you generate the app, okay? So instead of writing lots of code, you can get apps built like 10x faster because you're visually creating the app. And now we put AI into that so we can use generative AI to make it even better. So that's a big focus of us up the stack. So I think that -- there's a lot of other steps. For example, we put JavaScript directly in the database, the world's most popular programming language. There's all sorts of other stuff that we've done, but I think those are 2 really, really big deals. We think we have the best platform for developing enterprise apps with APEX and this new data model in the Oracle database.

Yes. And actually, just to add one thing to that, Juan. I mean, APEX, Application Express, I mean, it lets you build the application very quickly, as Juan described. But when the application is running, because it's being completely run inside the database, right, it's the database itself. It's the autonomous database, which is managing the application. There's no separate tier of the application that the customer then has to be for or think about scaling or pay for separately in any way. It's all built in. And I think that's probably, honestly, in many ways, at least as appealing as the ease of development.

Yes, ease of development. And the other thing is it's enterprise ready. You can build the world's biggest, most complex app in it. It's not a toy, a little development where you do this little thing on the side. This is like the world's biggest most complex application.

Right. As you mentioned, we're using APEX to build our own applications.

Yes. Pretty much every new application is built not only externally but for ourselves internally, is built on Application Express.

So before we close, are there any interesting customer examples that any of you want to share that would help demonstrate the unique value we're delivering?

[indiscernible] a customer of HeatWave, which is very excited to see, there's a company called [ EZ ]. You think of it as kind of like the Uber Eats of Latin America. I mean, they use HeatWave. They use a combination of the automatic machine learning that's built into HeatWave plus the Gen AI to do the following. So when you go and you place your order, or you're thinking about placing your order, they use auto ML to generate a recommender model, right, which then based on your past history comes up with a list of recommendations. They feed that list of recommendations into Gen AI to give you a natural language description of what they want to suggest you'd order, right? And that -- if you look at their application, the total amount of code they have the right to do that auto ML, the recommender engine, the Gen AI, it's almost nothing, right? I mean, if you look at the source code for these applications, it's very, very small. And I think -- and that's something which they weren't ML experts in any way, right? They didn't know anything about recommender systems. They didn't know anything about Gen AI. Yet they're able to build this very powerful application because of what we've built into that -- the data store.

We're enabling our customers to do so much in a much easier way, even if they don't have the knowledge. Any others, Juan or T.K., that you want to highlight?

I'll take one from -- again from health care. And this one -- I love this example. It's Advocate Atrium Health is a pretty big health care network. They love the integrated end-to-end offering that we have with health data intelligence, with the analytics and the closed-loop actions. But their entire -- they run Epic pretty much exclusively for their EHR right now. We'd love for them to move to Cerner, but they love our analytics so much so that -- and because we operate just like we're multi-cloud, we're also interoperable with all sorts of health care systems. So we've integrated well with the Epic system. So they're very happy to work with us, with health data intelligence and analytics that inter operates very well with Epic as their EHR. So I find that a pretty powerful statement.

Right. Open is the theme. And Juan, should we close with you on another example?

Yes. So we released our latest release just in May. So actually, we had a number of customer presentations here at CloudWorld and some of you saw some really big financial firms, retail firms, entertainment firms that are already adopting the AI technology and Oracle Database. So we had a number of these customer panels, and I've never seen adoption this fast. I mean, this is an enterprise database. And it's been, what, 3 months, I don't know, maybe 3.5 months. And we already have customers here talking about how they're using this stuff. And the reason for that is it's so easy. We've baked it into the Oracle data center. So like a 6-line SQL statement, you can combine all your business data and AI together and do amazing stuff that was never possible before. So this technology is super easy to adopt. And so that's well up people like, oh, I learned this in an hour. It used to be all this AI stuff. You basically had to go to school to get a data science degree. Now we have just regular Oracle Database developers just cranking this stuff out. So it's very easy. And yes, it's fantastic. We had a lot of customers showing what they were doing already here in CloudWorld.

That's great. Thank you so much. So much easier. We're expanding our reach. We're open. Thank you. T.K., Juan, Edward, I'm now going to let you guys go, and we're going to switch from database and analytics, and we're going to switch over to talk about applications. I want to invite up Steve Miranda, EVP of Oracle Application Development; Evan Goldberg, EVP of Oracle NetSuite; and Mike Sicilia, EVP of Oracle Global Industries. Let's talk about our applications business momentum now. Thanks for joining me. Steve, I'll actually start with you. Two of the many initiatives that you're working on, I wanted to pick on 2 to start with. One is the journey of fusion applications to OCI and to Autonomous. And then the second, of course, is going to be AI. We have to talk about how you've been embedding AI for many, many years into the Fusion apps. But let's start with the journey to cloud. Where are we on the journey? What are some of the benefits we've achieved? And looking forward, what's left? How are we going to continue to move to Autonomous?

Sure. Well, in some ways, we're done in that 100% of the Fusion customers are on OCI. And they've already achieved performance improvements. They've seen a reduction in terms of the time it takes to do an update. What they probably haven't seen or have a notice is increased security, behind the scenes in many different dimensions that OCI brings us. So from that sense, it's complete. However, in another sense, it will never be complete, and that's good news for our cloud-based customers. So an on-premise customer or somebody who chooses our competition, has a static cloud underneath them or, in fact, is on-premise. So what happens then is you build up technical debt, whether that's hardware, whether that's infrastructure, whether that's database, whether that's operating system, middle tier, I can go on and on. We don't have that because we are built and designed directly on top of OCI. So everything that Clay and Juan and Edward and T.K. talked about before, our apps just inherit that. So we update the database on a pretty continuous basis with the latest and greatest features. We update really every layer of the stack, both in speed, security, reliability. I'll save the AI part of the question later. The next big step on the database is obviously our move to Autonomous. So as Larry announced in his keynote yesterday, we'll be moving all of Fusion over to Autonomous starting in 2025. So we're well underway. Now that will give us another layer of performance security. And from our standpoint, it not only gets reliability and security, but also because it's fully autonomous, reduces the cost of us operating our cloud. And that will happen over time. And there's many dimensions, not only the people aspect, but also the scale-up, scale-down aspect that Autonomous gives us automatically to help us manage the cloud cost.

Right. It continues to enable us to achieve more with less.

Absolutely. Now the latest example then of what else could we get by sitting on top of OCI is AI. So again, our speed to react versus our other application customers is really because we sit on top of OCI. So as Clay has signed up essentially every LLM engine in the world, certainly the leading ones, we get direct access to it. And in fact, at one point, Larry asked me, well, which one are we using? And I just said, Larry, the answer is always the same. We use the API to OCI, which is to say we use the best LLM available on the market. We fully expect that's going to change over time. To our application customers, it doesn't really matter. We're going to have the best-in-class because that best-in-class is hosted on OCI. So what we announced last year was 50 AI use cases, things like using the LLM to generate job post, using the LLM to generate item description. Using it to summarize reports and financials, what we call narrative reporting. So we announced 50. We ended up delivering over 100 over the last year, and we showed those -- those are to all of our customers because all of our SaaS customers get those quarterly updates. So we way over delivered there. And then at this conference, we've taken the next step from, let's say, tactical uses of AI to more what we call AI agents. So these are more comprehensive use cases where you had an agent respond to questions about benefits, and give you clarifications there. And the important thing there is it's secure. So we never change any of our customer data to the LLM, it's contextual. So you don't have to tell the LLM who you are or where you are or what's your tenure, what your organization is because it's embedded in the applications. And it's also contextual because you're in a process that you're doing it. So it gives you a lot of efficiencies. We've added it to financials in a number of places, the general ledger being a notable agent, where you'll have most of what our financial analysts do are look for exceptions, ask questions, try to answer reports from executives. I think Juan talked about it earlier. Now in the applications, they're using English to do that. Give me the revenue trend of this account, show me the expense for this region over the last 6 quarters. Put that into a bar chart graph, so the executives can see it. You all can do it yourselves with ChatGPT today, and we've enabled the same in the application. So leverage on top of OCI in some ways, 100% done. In other ways to the benefit of our customers, it will never be done because we're on it. Everything Clay talks about, we push forward to the application customers.

So let's dive into the AIP. So you already started to hit on it. Can you give some of the specific examples of how you've been embedding AI into Fusion apps? And you talked about continuously releasing new features. I think it's hundreds every 90 days. Can you give us some examples?

Yes. So just as a refresher, because we have a SaaS application, these quarterly updates, every quarter, in every product area, so financial supply chain, HCM, CX release on average 100 features to all of our customers. About 80% of those are driven from customer requests. So we have an online forum that customers and partners, our system implementer partners give us ideas. Those ideas are debated, frankly, amongst the customers and partners, they're debated with our product managers and then they get delivered. And then, of course, there's ideas that we bring forth, which are usually combinations of technology and business, which is what AI has done. So this concept of agents are sort of full prospects. We've gone from tactical AI to agent AI, which is kind of full task that you can now embed with an AI. So I briefly mentioned the benefits case. Let me explain you how it works. So Europe company, Oracle, for example, we have a rather lengthy benefits document. It tells you what your health care rights are. And it does that per country because keep in mind, in the United States, you'd pick Kaiser UnitedHealth, but in the U.K., you have a public health option and every country is slightly different. Sometimes, there are regulatory rules around benefits. Sometimes based on your tenure, your vacation time varies or your maternity leave or leave of absence rules vary. So these are very complicated documents. Our customers and apps will load that document into a secure 23ai database that Juan talked about, we will index that. We'll use a vector search. And then we use the LLM on top of that. So in the context, when you're enrolling in benefits or you're asking anything in HR, you ask it these questions. I'm about to go on maternity or paternity leave. I have vacation balance left, kind of extend that. I'm traveling in Europe, do I have to do anything special to get a medical coverage? Or I'm in Europe on travel, I need a prescription. Is that covered? Things like that. Today at Oracle, we have people whose job it is to answer these questions. We took those sample questions. We applied it to LLM in this index document. The answers were, I'll say, excellent, and that's probably underselling it. Let's say, better than humans can do because these are enormously complicated documents. Not only answers the question, but it cites you exactly in the document where it got it from to drill down on. And so that's an example of an AI agent. This is not just helping you write text. It's a full-on process and we have 50-plus of those coming over the next year.

Wow, so enabling us to spend our time on the higher-value activities versus the more mundane tasks.

Now those benefits people can spend their time negotiating get us better benefits policies instead of answering questions like, hey, do I have to pay for my prescription.

So we're obviously a customer. I'm wondering if you can share some of the specific examples of how we're using AI in our Fusion apps internally.

Sure. Well, I mean, I'll just share the results, and we talked about this all the time, but I want to make sure that people are -- understand the importance of it. So we announced our earnings results, as you all know, on the ninth.

With a holiday in between.

Exactly, you stole my thunder. So that was our year-end close on -- or a quarter end close on a Saturday. Day 1 technically was a Sunday, day 2 was a holiday in the U.S. Then there was another weekend. So depending on how you want to count weekends and holidays as working days, it was 4 or 5, and that's not closing the books. That's filing to the SEC, that's having the fully audited. That's having the forecast ready and set up for you guys. Oh, by the way, we had this little event to prepare for along the way for it. So now in some ways, who cares that we announced it? But what it cares to some customers today or this week, given the analogy, it's sort of like driving in the fog. The fact that we can announce 9 days after and have that done, it's not done because we gathered all the data the day after the close and did that really fast. It's done because throughout the quarter, we have visibility. So if you're not able to announce until 10, 20, in fact, some of our apps competitors announced as much as 30 days after the end of their close, they're essentially like driving in the fog. They can't see that far ahead because they don't have those results. So what do we use AI for? We use AI for everything from character recognition on all of our documents to get us paperless, we use AI for audit and audit anomaly detection. We use AI for reporting to help again that narrative reporting part, really every step of the process we use AI. And the results are visible on a 9-day announcement. But what you don't see is that because we have the 9-day announcement, we have better transparency of our data throughout the year.

Right. And not only are others driving in the fog, but when you look at it across a year, it means we have an extra month or 2 that we're looking forward and they're still looking back.

We have an extra month or 2. Oh, by the way, you spend less time auditing your results. I mean, it's easy for the others, you spend less money that way. You spend less time with your people doing it. It's exactly what you said earlier. You can spend time doing value add to the business instead of sort of the logistics of the business.

And before I turn over to Evan into NetSuite, I just -- we talked about 2 of the many initiatives. Is there anything else that you want to highlight?

So the other big highlight is really, if you look at the move to SaaS, it very clearly started in terms of service companies, but we've really seen the momentum in the last couple of years moving to product company. And in fact, we had Zebra Technologies and DHL on stage with us that are full and not only financials, but supply chain. And that really speaks to the fact that we've really completed the move in terms of the cloud capabilities now far surpass anything we had in manufacturing in either E-Business Suite or JD Edwards. And then the big announcement we announced this year was something we call smart operations, which is really, I'll call it a modern MES application added to our manufacturing, which we never had before. And I call it modern because what we found from customers is that the days of what you classically think of, or I classically thought of an assembly line, you have workers and have got big gloves and an MES system with 4 buttons on it, those are diminishing. Now many, many, many assembly lines or manufacturing plants are part people, part robotics. And the people are dealing with fairly sophisticated instrumentation. So they have RF -- or QR codes or RFID and they scan that in their phones and they get work instructions on a tablet, and that's what they're doing. And so our smart operations with operators workbench, one, it's a tremendous advantage for our existing customers to be able to modernize; and two, it's yet another step, an incentive for those product on-premise manufacturing customers to move to the cloud, and we're getting great response.

Great. Thanks, Steve. Evan, I want to move over to you, and I want to start with a similar question as what I asked Steve because it's been a few years now since we acquired NetSuite. You've made a similar journey to OCI. You're now the largest ISV that runs on OCI. Can you explain some of the benefits we've seen as a result?

Sure. Well, I'd actually just like to cut and paste Steve's answer. No. I mean, very similar. And for our customers who are smaller, fast-growing companies, it's even more dramatic for them to be able to run on the flagship Oracle hardware, the same hardware that Fusion is running on, Exadata. NetSuite is running on. They're moving to Autonomous. We're moving to Autonomous and our customers massively benefit from Autonomous because they don't have the resources to tune the database. Obviously, they'll call us in and say something is not running fast enough, and we'll get a person on it. But now the idea that the sort of robots are doing it every night to make their NetSuite instance run as optimally as possible is fantastic. And of course, the same thing as with Steve, they're getting world-class reliability, security, performance. And again, it should be sort of transparent to them but we can measure it, and we can see the benefits that they're getting. So it's been a fantastic experience. And exactly the same as Steve said, all those services inside of OCI, we're increasingly taking advantage of. We use the machine vision service for our build capture, for example, and it works great, and it's fast and reliable and we can have -- I mean, we have 40,000 companies. So there's a lot. We pushed the infrastructure pretty hard, and it does not even bend much less break.

So I want to ask you about AI as well, the AIU continue to embed in NetSuite. But first, something that we do that's unique, I think, compared to others, across all of Oracle, not just in NetSuite is we're not charging for the AI that we're embedding. We don't have a smart version and a dumb version of our apps. Can you explain why we've taken that differentiated approach? And how does that translate into benefits both for our customers and for Oracle?

Yes. I think the reason we've taken that approach is because if you look at the very high level of what you want business applications to do is you want to be able to ask them a question, whoever you are in the organization, and you want to be able to get an answer back that you understand. And then you want to be able to have a conversation to drill down on different pieces. Exactly how you would work if you had like a world-class financial analyst at your beck and call to help you run your business effectively. That's the vision for AI and business applications. And why would we offer anything else. And I think that the dramatic improvements and insights that we can give to business entrepreneurs and everybody in the business, the dramatic improvements in productivity that we can give to them. We shouldn't -- I use the analogy, you wouldn't sell a car without wheels. I mean, you can't sell a business application without this stuff because it is absolutely central in what we're trying to do for the business user.

Right. It's like the difference between buying the Garmin GPS versus the GPS that's just embedded in our phones.

Exactly. I mean, that's the way the world is going. And we both -- Steve and I have a very similar philosophy of having a suite where everything works together, it's running off of a single data model. It's the best data to run AI on because you have comprehensive data about your business. If you're looking at a customer, you have comprehensive data about the customer. That's the way you're going to be able to sort of make the best inferences about like what's the best next step with this customer if you're in a prospect situation, what's the best next step if you want to upsell to the customer. That's one of the most exciting things we have going on. It's always been -- to me, the canonical use case for entrepreneur is how do I sell more? That's what we're thinking about all the time. So we've spent a lot of time even before the LLM revolution using traditional machine learning to help them sell more. But LLMs allowing us to make the interface better so that any salesperson can use it really easily, and it's super exciting stuff.

So let's talk about some of the specific examples. How have you been embedding the AI? What are some use cases?

Yes. So we did initially, I think, what a lot of applications did was just provide very easy access within the application to large language models. But I think we did it in a pretty unique way. We preconfigured every place in NetSuite where you have to -- where you enter text to automatically know about all the related data in NetSuite that's relevant. So if you're in a sales description, you don't have to say write me a sales description, it includes the price, and it also includes some of the features of the product. I mean, we preconfigured so they just say, write me a really peppy sales description, and that's it. And it knows how to bring in all the right data automatically. This year, we've introduced the fact that each business can actually configure how that works for every area in NetSuite. So if you have a particular brand style, you can edit the prompt in something called Prompt Studio to make sure that it adheres to your brand style. You can say there's some other custom information that we have that's really important, you can point it to that. You can choose even because the other is something Steve alluded to that the OCI generative AI service has basically every large language model available, you as a NetSuite customer can basically test and try each one and see which one does best and configure and maybe in one area, one LMM works better and in another area, doesn't. So we're giving a lot of power to the users. And then from -- again, this is something that customers won't necessarily perceive, but it's really, really important, their data never leaves OCI. And Oracle makes a promise to secure everybody's data, and this is another way doing that. So that's super exciting. And then I alluded to the sales, upsell how we call them intelligent item recommendations. Anywhere where you're dealing with a customer, whether you're in an opportunity, in a sales order, if you're just looking at the customer record, we're surfacing our assessment of what other things you might want to sell that customer. And we're using all kinds of data. I mean, we have very rich data in NetSuite. We're using everything we know about that customer. What region are they in. What size are they -- anything we know to make these recommendations really intelligent, and we have metrics that show that it's working. In the last quarter, these recommendations had an 11% conversion rate. Customers use them to do an additional $5 million of sales, and we're just getting started. But again, entrepreneurs, they want to sell. They love that.

Yes. Before I move over to Mike, is there anything else you want to highlight that -- and you want to share with the audience around how you continue to expand NetSuite's reach and as well deepen and expand the relationships with your customers?

Yes. So Steve and I actually, and our teams, have been collaborating on the next-generation business user powered by AI. I mean, AI has so much power just to make the experience better. You can recommend ways to use the system. Obviously, you can give this incredible next-generation help that's tailored to how -- what you've learned about how that -- what kind of learning level that user has. So AI as an important component, but also just making these interfaces to business applications, which have traditionally been horrid. You come in on Monday morning and it's like, ugh, I got to log into my business application. With the team that is working on Fusion and NetSuite, the Redwood design system, we really strive to have users fall in love with these business applications one interaction at a time. And that may sound unrealistic, but we're getting incredible feedback that they're feeling like, wow, this is much more like the applications that I'm used to using and that I like using in my personal life. And I've never seen a business application that looks and feels like this. And so that's a really exciting area for us.

Yes. It's not just pleasant. You're actually guiding each employee to do their best work.

Absolutely.

Which I think is great. So Mike, I want to turn it over to you. We just talked about horizontal applications. Your team is providing the mission critical applications that our customers need across industries. Can you start by explaining how are you seeing customers leaning in now maybe more than they were 6 months ago or a year ago into the cloud? And what do you see for the future?

Yes. So the mission-critical applications, particularly in heavy regulated industries like banks and health care, telecommunications, utilities, it's not unusual that if you look at the data centers -- the customer data center today, you'll find a lot of mainframe. You'll find a lot of AS400s that are running these mission-critical systems. And we -- there are cars that are now classic cars that are younger than the age of some of infrastructure. And I think what's really -- there's a couple of things that has changed the mindset and think about how do we move that mission-critical stuff to the cloud? The first is the success they've had with the backlog, right? The success that they've had with it, the quarterly updates, the security as a service, the innovation as a service. And there's now a comfort level to say, well, actually, we need to think about moving all of it. It's becoming very hard, very expensive to maintain old code. And if we don't start to modernize it, we potentially have business impacts. We also potentially have major cybersecurity impacts as well because it's also very hard to secure that stuff because at some point, there's some there's some Internet-facing gateway into that. So that's what I think you're seeing in the heavy regulated industries. So again, banks, utilities, health care, telcos. Another class of industries like retail, hospitality, food and beverage. You heard in Safra's keynote yesterday, again, in Mike, I had 3 customers, all of whom we talked about the benefit Cloud for omnichannel transformation. So it's really hard to do that by yourself. And just with the changing nature of the business in retail and hospitality and food and beverage and everything from delivery -- third-party delivery aggregation to customer intelligence to upsell wheels, and you heard about that from MGM. It would be difficult to sort of stitch that together on your own. So I think that those drivers are really getting people to start to think about time to take the mission-critical applications to the cloud. I would say very good progress in mission-critical applications to the cloud in retail, hospitality, food and beverage, merchandising planning systems, order management systems, payment systems and so on. And the bigger regulated industries like banking, what you saw a couple of years ago was starting to surround the mission-critical application with cloud services like financial crime and compliance, anti-money laundering. Now the conversations are about moving accounting foundation services to the cloud together with ERP, together with Fusion in our public safety businesses, moving mission-critical tactical response some that cloud together with NetSuite as well. And it's really, I think, that push and pull. They've been very successful in the back office. They love the momentum, regulation, hard to keep pace with regulations because you've got to certify this old stuff and it's actually compliant. And now it's time to move the back -- the heavy duty, in some cases, very heavy iron to the cloud. And I think that's -- I don't think that there's a category that I can think of, at least in the 10 vertical industries where we supply in an automation where somebody has said to me, clouds a bad idea on clouds off the table. That's that -- 3 years ago in certain verticals that may have been a conversation, that's no longer happening.

Definitely accelerating momentum. And you just hit on our industry cloud application together with Fusion and NetSuite. But you're really bringing not just our horizontal and vertical applications together, but database and OCI. Can you talk about how you're delivering really holistic end-to-end solutions to each industry?

Yes. There's a couple of things that I think that are important to highlight here. The first is not only are we pre-integrating Fusion applications, let's say, with accounting foundations and for banking. We're also creating vertical-specific functionality inside our back-office application. So Steve's got specific functionality in HCM and supply chain management for health care. Not a custom version, not a customization, off-the-shelf, available features that are applicable organizations. In NetSuite, we have a specific version for local government to deal with payroll management and things like that. Again, not so not a custom extension, this is off-the-shelf software. So bringing these 2 things together from an integration perspective and then further tailoring for the specific needs of the vertical industry, I think, has been something that customers are very, very excited about. The other thing is when you talk about end-to-end solutions, like about front-office applications, back-office applications, analytics and OCI is the infrastructure. The conversation with customers is about outcomes. And I'll give you a couple of examples. In health care, the conversation and one of the buyers is the Chief Medical Officer of health care organizations. And I can talk about the performance, scalability and security of OCI. I can talk about our IoT aggregation network. I can talk about our ability to deliver private 5G networks in the hospital using our enterprise communications platform. I could talk about our ability to arbitrage low-orbiting satellite, terrestrial networks and cellular networks so that you've got triplicate coverage in a mission-critical operation. A much better conversation is to say, in your ICU of the 32 beds, which patients are likely to progress to a septic situation within the next hour? That's an outcome. It's very difficult to talk about if you can't supply all of those things. Now arguably, they can go to a bunch of different vendors and stitch that together on their own. But if we can supply that as a service, we can supply it as a HIPAA compliance service. And we can supply that with a quarterly update and cyber defenses, which is a huge struggle for health care. Today, it's a very -- it's a highly differentiated conversation and that's exactly the conversations that we're having, right? Because it's all of that together that allows us to have instance telemetry into something like an ICU, you can extrapolate from there a bunch of different use cases in health care. In our public safety business. which we launched a couple of years back. We had a big demo here today. This week at CloudWorld with the cyber truck in the center of the floor, lots of buzz, lots of activity about all the technology there. The outcome-based discussion because of the OCI computer vision, right, because of our APEX low-code generation activities and because of the Command suite that we put together in vehicle and in vehicle on first responder and more things. The outcome discussions about is about weapons detection and tactical response generation. The reason we can have that conversation is because using OCI computer vision, we can process 30 frames per second. Nobody needs to know the technical detail. What they need to know is that I can instantly identify and hopefully neutralize a threat, far much further away from the perimeter of a facility to something's trying to breach, right, with fixed asset cameras, IoT networks and integration. So I think that it's just something that's very unique for Oracle, right, is that we're having that conversation based upon what you -- what's the transformation that you're trying to achieve? And it's because we have all of it.

It's not just unique value, but we're really delivering new value.

Absolutely.

I didn't ask you the question about your cloud applications moving to OCI because I think you'd give the same answer as we've already heard. But I will ask you about AI because I think -- I can't let you leave the stage without doing that. How are you embedding AI into our industry cloud applications? What are some use cases?

I think in a few stages. The first stage, which is now widely accepted, and we've gotten terrific feedback around is what I'll say is because these industry applications are often at the edge, right? It's automating routine tasks that are very laborious, very manually intensive and people get tired of doing them. In health care, for example, doctors, nurses, there's a lot of paperwork generation. There's a lot of documentation. And this is the perfect vertical industry where AI essentially is the UI, right? So instead of having an electronic health record system, where you've got to click through 10, 12, 15 screens, the process of patient interaction. And you got to do that while you're trying to be able the patient at the same time. If we do not -- if you do nothing, but let the autonomous system listen in the background, semantically break down the conversation, automate the ordering of labs, automate the ordering of prescriptions, automate the document process. That's a pretty big win. That's generally available. We released that in June of this year to just terrific accolades from our customers. So that, I think, is the first example of applied AI. You can imagine in every vertical industry that there are these edge use cases where they're highly repetitive tasks that are -- don't have a whole value add when you're trying to transform patient experience or customer experience. The next generation of it is what I spoke about with as the example of continuous telemetry into a cardiac -- continuous cardiac monitor, whereas you need machine learning, you need low latency network, you need to be able to figure out what the patterns to be able to make predictions about what may happen to a patient. Predictions about where a customer may choose to take their energy business in deregulated markets and predictions about where customers may choose to shop in an omnichannel environment for retailers. And then the sort of the next frontier, if you will, right, is things like molecular discovery. And looking at our clinical trials business and -- our pharma customers are saying, look, if we can really access these GPU clusters at scale, and we can look at figuring out doing much of the simulation using computers for new medications and molecular discovery, we can not only save a lot of time and a lot of money but potentially develop things that are far more safe and far more efficacious. So I could go on and on and on about each vertical industry and give you a very deep answer on all of it, but hopefully, that gives you a flavor of, I think, the layers of AI that are -- that consumers are clamoring to.

Yes. The theme of it is we're enabling our customers to automate their operations not only so they can save money, but also so they can focus on their missions and advance those. We're almost out of time. We've talked a lot about our business momentum and applications. Is there anything else that you wanted to close on for the audience here?

Well, hopefully, you got a chance to stay during the week because I would say that -- say what we will up here, but the excitement and the buzz and the crowd was very, very noticeable this year, and I felt it in all my customer meetings, I felt in all the sessions that I attended. So the excitement, more importantly than us saying that the excitement from our customers to receive it and really looking for adoption was -- you could feel it in the buzz in the room, in every room that I've been in. So I thought it was just great.

Great. Evan, how about you?

Yes. I mean -- well, I think the thing to highlight is the collaboration that we have across Oracle. Between Fusion and NetSuite, we cover basically every sized business, and we have all that amazing business data. And a lot of the things that -- the problems that we need to -- Steve and I need to solve are similar, and we're increasingly collaborating on that. We're taking some of Steve's amazing products that were aimed at enterprise, and we're scaling them down for our fastest-growing NetSuite customers, and they are eating it up. So just whenever you do these sort of acquisitions, it takes a while to really get in your groove. And I think we are now just working incredibly well together.

Incredibly. And Mike, how about you?

Well, look, I always have a lot to say, but I think what's far more important is what our customers say. And what I continue to notice the trend is how many customers are not just talking about the technology, but are talking about business transformation, clinical transformation. And I said in my keynote yesterday and learned a lot, even in the keynote from our customers, just about how the impact of thinking about this from one Oracle, from an end-to-end strategy. So to me, that's incredibly exciting. As Steve and Evan said, the buzz has been terrific, and I'm quite happy with for our customers and humbled by what our customers have been able to do with our technology.

Great. Ultimately, our success is all about our customer success. So thank you for joining me. Thank you, everyone. I really appreciated the opportunity to talk about our business momentum. And with that, we'll move on to the next session. Thanks.

Please welcome to the stage Oracle Executive Vice President, Jason Maynard. He is joined by Eash Chitimalla from MGM Resorts International, Hiro Hamada-san from Nomura Research Institute and Pedro Sardo from Vodafone.

All right. We got that Clay energy flowing, I can tell, right before lunch. We've got the most exciting panel. All right. I think before we get started, I've been instructed to alert you to my favorite slide. I missed this slide, I'm not going to lie, the disclosure slide right behind me. All right. So thank you guys very much for coming. But really, I want to thank our 3 guests here. Like I said, I have the best job of the day, which is I get to share with you all some of our great customer stories. And you can learn a little bit about what they're doing with Oracle products. And what's really exciting to me is we brought customers from all around the world. So we've got customers here in Las Vegas, Tokyo, London. So you'll get a nice global flavor as well across the entire product portfolio. So first, let's kick it off a little bit. Pedro, why don't we start with you? Why don't we talk -- tell everyone a little bit about yourself and the work we're doing together.

Okay. So -- hello, everybody, and thank you for inviting me. I work for Vodafone. For the ones that do not know, Vodafone, we are a mobile operator that mainly covers Europe and Africa. Again, we provide a lot of typical services that you'd expect from a telco, fixed, mobile, TV, IoT, and we have around 300 million customers. Within Vodafone, my responsibilities are I manage all of our operations and data centers, which includes all of the infrastructure, and that's why we have also -- one of the reasons why we have a close relationship with Oracle. And because we have a very strong in-sourcing strategy, I also manage our technology centers that act a bit like an internal SI that work exclusively for Vodafone, where we have already relatively good size of 15,000 people working across operations, delivery. And so a lot of the projects that we do together, again, instead of sometimes using SIs, we work together directly with companies like Oracle.

That's great. Thank you. Hamada-san, maybe tell folks a little bit about yourself.

Sure. My name is Hiro Hamada. Thank you for having me. I'm from NRI. NRI is a management consulting and IT services company, headquartered in Japan, and NRI has a long successful history of partnering with Oracle. As a matter of fact, NRI was the first in the world to adopt the OCC. So now NRI operates OCI in our data centers in Japan. And also, we recently launched Alloy as well. And NRI -- one of NRI's core business is its first solutions for financial institutions. They are widely used in Japan, accounting for about half the trading volume on the Tokyo Stock Exchange. There are multiple solutions covering both buy side and sell side. And we have migrated all these solutions to OCI. So they are absolutely mission-critical applications, and they are all running on OCI.

Fantastic. Thank you. Eash, yourself?

Thank you for including me here. At MGM, our mission is to entertain the human race. And Oracle is a key partner. We use a number of core systems for hospitality where we check in the guests, check out the guests using Opera. We use the financial systems where we close our books using Oracle Fusion Financials, and we use the supply chain systems to do our forecasting, inventory management and how do we make sure our warehouses are managed.

It's fantastic. Let's go back and let's dive in a little bit here because we obviously have had a lot of news this week about not just multi-cloud, but obviously with the database. So, Pedro, let's kick it off and talk a little bit about how you're one of the first multi-cloud customers with Oracle database at Azure. Talk us through a little bit about what your thought process was in making that move and how you're thinking about this multi-cloud strategy.

Look, maybe let me start a couple of years back because I don't think that even for us multi-cloud was an obvious thing in the beginning. We started using cloud probably over 10 years ago. And that for several years, we tried to keep into one cloud, okay? And at that point in time, it was more or less obvious which one it was. And for -- we tried for multiple reasons to stay with one cloud for standardization, consolidation. But quite quickly, and that was probably 6 or 7 years ago, we got to the conclusion that, that was the wrong strategy. So we actually decided to go multi-cloud over 6 or 7 years ago. Probably our first big partnership that we did on that part was with Oracle -- sorry, it was with GCP in terms of analytics, because when we decided to go multi-cloud, we also needed to give a lot of guidance to our teams, okay, so which clouds are we going to use for which workloads. And so the first one, big one was then with GCP for all the analytics workloads. And then we have been expanding again. Then we did also a lot of work with Oracle. We started actually first with OCI, our first solution with OCI that managed our retail stores in the U.K.. Moved to OCI over 5 years ago. And then probably 3 years ago, we started discussing about DRCC, okay? And the reason for that was that we were finding out that we're pretty good at migrating all of our and creating all of our workloads on public cloud, especially when we are talking about new applications, but we were getting a bit behind when we were talking about our more core applications that were probably -- I don't like to use the terminology of legacy because some of them, they are really critical and cover our core processes. We were not doing such a great job of modernizing those ones. So that's when, again, the partnership with Oracle came. Again, we created 6 DRCCs together with Oracle. And ever since then, we have been migrating applications and databases, and we have been modernizing a lot of our databases into DRCC. Then we also started discussions with Microsoft because there was a need for -- to have some -- certain workloads in Azure. And we also signed a partnership with Microsoft. I strongly believe because when we were discussing the partnership with Microsoft and how big that was going to be that probably the partnership between Azure and -- between Oracle and Microsoft did not happen, maybe our partnership with Microsoft may not have been so big because the Oracle database is a really key application for us. It's a really -- most of our core systems rely on the Oracle database, and we plan to continue to rely on the Oracle database. So if I try to summarize the ability of us moving the workloads to where we need to move to because of whatever reason it might be and always have such a component, such a critical component of our architecture, which is the database, that is Oracle, and have that available into all of them, I think for us it's a major surprise, to be fair. Probably 12 months ago, we were not expecting to have that benefit.

No, it's great. It's interesting to hear you talk about the flexibility and the choice and the ability to run in a heterogeneous environment. What are some of the benefits that you're seeing and expect to see as part of this move?

I think that the key one is really the flexibility and the optionality. Let me give an example. When we started working with DRCC, one of the reasons why we put DRCC was to enable us to do modernization layer by layer. So I might be modernizing the database without having to modernize the application. And then later on, I modernize the application. DRCC, being on-prem, being close to wherever the application is running, it enables us to different pieces of the application and modernize them at different places, that is optionality. Later on, if we want, we can then flip all of those applications when they are all modernized into public OCI. We expect to sign about this -- the benefit of having Oracle at Azure, at GCP and now at AWS, enable us to -- gives us optionality. We can then choose what we want to do each one of our applications.

No, that's great.

We just signed -- again, we are in the beginning because we signed the agreement 2 weeks ago. Although we have a longstanding relationship, this specific agreement, again, is pretty recent.

Yes, this is just getting rolling, but we'll have you back next year, we can talk a little bit more about it, so that will be great. Hamad-san, it was interesting, flexibility, optionality. You're building an AI platform on Alloy, which was really designed in many ways to enable from a business model standpoint that flexibility and optionality. What are you looking to achieve with that? And how are some of your customers going to take advantage of this AI platform that you're delivering?

Yes. So let me give you some background. So I think the financial industry has very high expectations on GenAI, and many POCs have been already completed. And the real value lies in how to integrate AI with the existing core systems; however, there are challenges. Much of the data that financial institutions handle include confidential information such as PII, so they need a secure and robust environment to fully leverage the power of AI. So NRI's answer for this need is our financial AI platform. By fully leveraging the latest technology that OCI provides, we provide the secure and robust environment that our clients need. And I think this platform will empower the financial institutions to transform their existing core systems through the power of AI.

No, that's great. I think everybody in this room, obviously, working in the financial services industry can appreciate security and privacy of data and the importance of that. So that's fantastic. Eash, we're going to shift gears a little bit and talk about applications, right? And MGM Resorts, obviously a fusion customer, as you mentioned, with ERP, EPM, supply chain, but you're also using our OPERA hospitality platform. OPERA is moving to the cloud. You have a lot going on in terms of a modern transformation. Give us a little background and discussion of how you thought through this evolution and really what you're looking to achieve from this transformation.

Yes. As we go through upgrade cycles for each of the core systems, one of the things we looked at was partnering with Oracle understanding what does the cloud bring, what is the transformation that we can achieve through the cloud and then purposefully made the decision to move to cloud for all of these systems. There are a few key advantages that stood out as we did that analysis. One was efficiency. As we go to the cloud, our team can be hyper-focused on their business needs and not have to worry about maintaining the infrastructure, the cycles of upgrades and work through the maintenance cycle that typically drains out some of our focus. The second is security. Security is very, very paramount for us. We have confidential data in all of these systems. Moving to cloud gives us advantage, taking off what Oracle brings to the cloud maintenance, but also the cloud systems are built more secure. They are more flexible from access controls to data retention policies. That allows us to keep our systems more secure. The third advantage is employee experience. All of our employees use these tools day in, day out to support our guests. As we go to the cloud, the modern UI, intuitive UI allows the employees to be more efficient and get through their tasks very, very quickly. The fourth is around extensibility. Oracle team has done an amazing job in redesigning the applications to be cloud-native first. So there's a robust set of APIs. There's more open protocols used. That allows us to extend the systems very easily and build integrations. And the fifth is around future-proofing ourselves. As we go to cloud, we are on a faster cycle, we are able to release capabilities faster to the business. And we are also able to take advantage of additional modules that are in the cloud, get all of our data into one place, so we are able to run reports and decision-making processes faster in one system.

Well, I have to follow up on that one because I think it's -- that's such an important and interesting point, which is you're taking the entire suite, consolidating data, and in every business we're in, it's the fragmentation of data that drives us crazy, right, trying to get answers to questions. What are some of the things that you've seen as a result of being able to have that easy look across the business to make those decisions?

Yes. We are still working through all of those processes. We're not at the end state yet. But some of the advantages we have seen is our reporting cycles are faster. As you mentioned, consolidating the data into an external data warehouse and then compiling the analytics data takes time. One big advantage is near real-time analytics and the data that comes out of it to help our operational leaders make the decisions faster. We have seen how our operations with hotel, finance and revenue management are all able to take some advantages as we are on that journey in making those decisions faster.

No, it's great. And MGM has such a great reputation for customer service and care and bringing great offers and capabilities. It's -- you can -- it's so cool to hear about some of the work you guys are doing there. So -- all right, now as you guys all know, we're mandated at this point to talk about AI, right? There's like a quota on the AI conversations that we have to have. And these guys out here, they're all like tell us about AI, what's going on with AI. So we all know you guys are doing some stuff in AI. But I think maybe let's start, Pedro, to go back to you. Give us a few minutes, just talk about how you think about AI in the broad context of your business? And then maybe give a couple of specific examples and ways that you're starting to get rolling with it.

Okay. Good. And let me also give a bit of background on that one because I was mentioning that we -- one of the first partnerships that we signed was actually with Google in terms of analytics. And that's when we decided that we are going to have a very simple strategy regarding to data, which is our data was only going to reside in 2 places, either on the operational system, where, of course, it's mandated so that the system can operate and our -- on our [indiscernible]. Probably at that point in time, we did not foresee how important this -- ensuring that all of our data, which is in one place, was going to be so important for our AI initiatives. But, look, fortunately, that started. And today, we have over 20,000 terabytes, which we think is around 70% of all of the data in terms of analytics that we want to roll into our automation already there. We have 10,000 pipelines moving those data. We have already 600 AI models running. So I think that is how we started. We also wanted -- I'm not sure if everybody understands that. For instance, in Europe, GDPR is quite strong, and there are some variations according to countries. So we wanted our engineers to focus on how they develop those and explore those data and those AI models instead of being worried about anonymization of data, privacy, security and so on. So we created a platform that enables all of the development and operations of all of our models that is completely transparent to the engineers and to the analysts that are doing the work. We call it AI Booster, but it's basically based on the Vertex product from Google. And that enables us to really -- it's a massive accelerator on how we can use and how we can build models. When GenAI started to come, we were also very clear in terms of our strategy. Again, we are not going to build a new LLM, okay? But what we've built was an architecture that makes us independent from the different LLMs and enable each one of our analysts and -- to select which LLM applies better to their need. And -- all of this, again, just enables the people to focus on what they -- on the use cases that they need to build. In terms of what we see, in terms of the use cases for AI, I will probably -- we see AI at this point in time as probably, I would say, falling into 2 fields. One is augmentation of -- and I'll give a couple of examples, augmentation of the work that we all need to do or sometimes to enable things that will not ever be possible in any other way. When you talk augmentation, again, we all use Copilot 365, for instance, to help us to write a new mail in a better way. But again, Copilot can help our sales teams to better build a proposal, can help our call center operators to respond better to our customers or can help, again, even our software engineers when they are developing code to accelerate and to be faster and more productive when doing code. That is one clear example. And the second one, in terms of the things that probably will not be possible. One example that I think has been very successful for us is we -- every day, we take all of the calls that we receive in our call centers, we transcribe them to text, and then, we ask a summarization of all of the calls that we had the day before. And that is extra, that is giving us some insights, early insights sometimes into problems, where we can start seeing that, that some problems might be appearing that in any other way, it would be impossible to pick it up. So that is an example of something that is really probably not possible without the summarization capabilities that we get out of AI. Of course, then we have the chatbots. But again, I think that one -- I do not even mention that one because I think that, that one is the one that everybody traditionally expects.

Yes. No, I think that's what's so interesting right now, though, is that you're seeing all sorts of different ways that organizations are thinking about AI from routine business processes. I think what you're talking about here, so it's going to be a very interesting evolution in terms of bringing AI to your data, running your Oracle database and then actually even running that in multiple clouds. So this heterogeneous nature of what your existing IT environment is like and then being able to put your Oracle data where it needs to be is going to open up a lot of opportunities, so it's really beneficial.

Even how we deploy network, even the way that we deploy network, where we decide to deploy capacity, we use a lot of AI models to help us and to help the engineers decide, okay, I'm getting much more profit if I deployed more capacity in this location versus this location.

Yes. Hamad-san, you're literally jumping right in with this AI platform. You're like I'm -- we're going. Give us a little bit of the how and the why behind the scenes in terms of what motivated you and where you really saw the opportunity?

Yes. So I think, as you mentioned, data is really the key to successfully use AI. So -- and I think for every company, existing application is a great source of data. So it doesn't have to be limited to existing applications, but it's a precious resource of data. And one thing we did recently is preparing for a demo application for this chatbot to showcase. So what we did was -- so we built a sales assistant chatbot designed for entry-level financial advisers. So this application highlights how we can improve the effectiveness of the entry-level workforce while staying compliant. So I think this application is a prime example of how AI can be seamlessly integrated into existing applications.

I can appreciate that use case, lots of salespeople in trying to make them smarter and do better and take care of customers. That is exciting. That's great stuff. Eash, you're using Fusion, which has embedded AI. Maybe what are you thinking about bigger picture with AI and some of the things that you're hoping to achieve?

Yes. I think AI is so interesting, and as Hiro said, this all boils down to data, what quality of data do we have. For MGM, the way we are thinking about AI is focusing on deeper understanding of the guest, how do we understand the preferences, insights through the guest journey. We have so many touch points that we implicitly, explicitly understand the guest, how do we capture all of that to build a very deeper understanding of the guest and personalize their experiences? The second focus area is around operational efficiency, both in technology and on the business side. In technology, we use AI throughout the life cycle, from engineering, as we already discussed, as well as our product, QA and our operations, understanding where systems might likely to fail, where we have vulnerabilities detecting those through AI-based tools. And on the business side, getting the data faster, something that we touched earlier on, and allowing the business the key insights. And one part that we are fully working with Oracle is understanding the Fusion Data Intelligence and use that to get the insights to the business and operations team faster. And the third piece that we are looking at AI is the guest interactions. So we did -- like I think most people have, we launched a chatbot as well for the guest interaction to allow self-service while during their stay. And we are now looking to expand that to their trip planning and post-trip engagement.

That's great. It's exciting. I think the work that the teams are doing with Fusion and analytics and data and bringing it together in Data Intelligence will be really exciting for your -- for the back office, but the front office is really going to be cool. That customer interaction piece will be exciting to do that. Let's stay with you for a second here. We've got a few minutes left. What's kind of next? What's on the horizon? What are you working on? What are some of the things you can talk about?

Yes. OPERA Cloud is the biggest, probably the project for us. It's a massive uplift, rolling out to all of our 16 properties. And each property has a unique brand. And one thing we really are aiming as a foundational is to standardize everything and allow the properties to express their unique brand within the configuration options and not customize the core platform, and that's a core mission as we migrate out of our on-prem to OPERA cloud. And that's one of our big project next couple of years.

Well, and it's going to be really exciting to see the work you guys do because in '27, we're going to be moving CloudWorld to MGM. And so we're all going to get to benefit from this. So all of you in the crowd, you're going to be like, yes, I heard it here first, and you're going to have a great experience.

We will check you in through OPERA Cloud at that time.

You'll check us in. It will be amazing. I won't have a blow dryer in my room or anything like that. I'll have the nice pillows. You're going to customize the whole thing. I can't wait. I can't wait. This will be great. Hamada-san, what's next? You got a lot on your plate here with AI, but where are you going to go? What's in the future?

Definitely. So it's going to be twofold. So one is we would like to leverage AI technology as much as possible through our system development work, so coding, testing. We'd like to make it as efficient as possible. And the other part is we would like to keep our clients' new value or additional value leveraging AI. So I think the next immediate step would be building something that we can showcase how our clients can leverage that GenAI technology.

No, that's great. You're going to be building more cool stuff and making things better for everyone. That will be great. Pedro, we're going to give you the last word here with a couple of minutes. You got a lot on your plate, I know, so tell us what you got.

Of course AI.

AI. There you go.

But now on a more serious note, yes, of course, in AI, we need to -- we are kind of experimenting and trying to really understand what is the real benefit that we can get from all of the use cases that we are doing. And if I focus on a lot of the work that we are doing with Oracle, we have all of our cloud migrations that we want to execute. We want to modernize our infrastructure and our applications. We have a big initiative in terms of how do we take the good things that we see in some of our local markets that have taken some of the Oracle applications and make it into a state-of-the-art stack that serves our core business, that has the agility, the time to market, that enables us to deliver faster products and services to the market, how we can take that to all of our markets, because, again, having that only in one of our markets is not enough. So now what we are looking is how do we take all of that experience of one market that gives us the agility because that's what our business is asking, we need to be faster in taking things to market, how we do that and be able to leverage it across all of our markets.

No, this is great. I love the themes up here, bringing AI to data, run your Oracle database in any cloud, flexibility, choice, security, making sure we take care of that information and then enabling these end-to-end transformations, whether you're building applications, you're using Fusion across the board. It's great stuff you guys are doing. So I really do appreciate you taking your time, and we value our relationships with all of you, and it's just fantastic. So for myself and everyone at Oracle, we really do appreciate. Thank you for coming to us today. And everybody, thank you for the time we can share the stories. All right. Thank you all.

We will now take a short break. Lunch is located in the hallway. Our program will resume properly at 12:30. [Break]

Please welcome Doug Kehring to the stage.

Hi, everybody. It's great to be back again this year at the Financial Analyst Meeting. I'm Doug Kehring. I run operations here at Oracle. Many of you may have seen me last year. But here I am again. This is one of the better weeks that we ever had, given the fact that we had an excellent earnings announcement on Monday, we've got financial analyst meeting today, and in between it, we had all of our favorite customers and partners who joined us. So my goal today is to translate the product and customer discussions that we've had earlier in the morning into how it's impacting our financials. So as a result of this, you're going to see me being very, very cognizant of my notes because Safra said, don't screw this up. So just to be very blunt, I want to be careful about what happens up here. So as you've seen, we've got our safe harbor language, we also are going to use non-GAAP measures in this presentation. So here's the requisite disclosure again. And then finally, we've got the fact that this is -- there may be some future product direction discussions in here as well. So this is for informational purposes only. Okay. Let's get started. So as we reported during earnings, RPO is now greater than $99 billion. It's up 52% year-over-year. This figure has grown rapidly and clearly shows the pent-up interest and demand in the Oracle Cloud, which is now almost 3/4 of the total RPO amount and is up 80% year-over-year, actually over 80%. This figure really sets the stage for our discussion today because it's all about satisfying this amazing demand by delivering profitable growth. Now of course, RPO grows when the commitments our customers make to Oracle grow. This RPO figure and its absolute size points to the multidimensional interest from our customers. As we talked about earlier today with all of the great development leaders that were up here, it's across our entire portfolio from applications to analytics, from database to infrastructure. But beyond that, it's across all geographies and across all industries. And it's not just across our traditional enterprise customer base, but it's across many different customer types that I'm going to highlight today. The demand is widespread, and it's been accelerating. Now the historical RPO size is driving our ability to accelerate revenue growth. And as we forecast last quarter and we confirmed again on this week's earnings call, we expect to deliver at least 10% revenue growth this fiscal year. Using our previously committed FY '26 revenue target of $65 billion, the implied growth rate for FY '26 is 12%. Now not only are we confident in our ability to meet this target, but we expect to exceed it. I've got even better news to report on this as it relates to our future. But I'm going to save it for the end. So unfortunately, you're going to have to put up with a few more minutes with me before you get to hear the good stuff. So there are really three drivers behind the confidence in our ability to deliver. The first, as we've been saying, is we believe we've built the best technologies for the cloud in the world. The second is our product differentiation is driving increasing wins and momentum. And the third is we're building the capacity in order to convert this momentum into accelerating revenue and profits. What I'm going to do today is drill into each one of these areas. So let's start with our product and service differentiation. Unfortunately, I can't do justice to what the development guys were talking about today, but I'm going to sort of wrap it in a little boat. First, our product portfolio has never been more comprehensive or more complete. And as you heard, it's completely AI-enabled. And as we know, AI is transforming everything. And Oracle is poised uniquely to help any organization with that transformation journey, whether that's about transforming their data in order to use our end-to-end database features to ensure that they have the best and most up-to-date data to drive their automation or it's about transforming their line of business functions. Whether that's front office, back office, supply chain, HR, whether you're a big business or a small business, we have it all to help our customers make their employees' jobs go even better. And we transform their business processes. So using AI and data combined with our applications allows us to help transform customer experiences, employee experiences or any other experiences where our customers want to deliver better outcomes at a lower cost. And finally, we're transforming entire industries. This is really the Holy Grail. It's using the power of all of our technologies put together where we can help our customers achieve just that. We don't think anyone else out there can do it. And I was about to go through a whole explanation of how -- what that means, but I just realized that when Mike Sicilia was up here, when he talked about health care, that really resonated with me. That was him discussing the power of all our technologies brought together. It's not just in health care but many other industries where we're doing that with our customers. So now beyond that, all of this is connected together -- sorry, I apologize. That's part of the health care thing. Let me get to the second area, which is we add to that portfolio, the most flexibility, so customers can get the cloud delivered to them in a manner and at the price point that they want. Our data center or yours, the customer, public or private, sovereign or security air-gapped managed by us, managed by our customers or managed by partners on OCI, on Azure, on Google Cloud or on AWS, we do it all with the same cloud capabilities and consistent pricing. One stack for all possibilities. And we can do it because we've built OCI differently. Clay talked a little bit about that, a lot about it, but let me highlight one very interesting point, which is our cloud infrastructure can start with all of our services with just three racks. It's a fraction of the size of everyone else. In fact, we believe the minimum footprint of our competitors is over 140x larger to get started compared to what Oracle offers. This differentiation and flexibility is exactly what customers want, and that's what we talked a lot about this week. It's why we're becoming so popular in the cloud. And then finally, we enhance our technologies by activating our customer success services organization for every customer. As Safra has repeatedly mentioned, customer success is now at the heart of everything we do. We've built a customer success journey to accompany the customer from the start of the buying decision through their ongoing usage, making sure that they get value from every dollar of investment that they make with Oracle. It starts with helping our customers as soon as they make a buying decision. We guide customers and train their users on how to be cloud-ready. We then work with our implementation partners to ensure that our customers get a smooth and successful go-live. Then on an ongoing basis, we work with our customers to operate the cloud from proactive issue resolution to helping them drive value realization. And finally, we assist them with evaluating and deploying new features and innovations such as AI as it becomes available. We're also in a unique position as a result of this to help guide them to the next best Oracle product and service that they can purchase. We take the best technology, the most flexibility and an obsession with customer success, and it's this combination that's driving our cloud momentum with customers. Let's move on to the second driver. It's how we're converting these product and service advantages that I just went through into customer wins and momentum. So let's go back to RPO because we really enjoy this number. This velocity has been accelerating. It's moved from a 12% CAGR a few years ago to now 46% over the last 2 years. It's obviously being driven by the cloud. As you can see, cloud now represents 72% of RPO, up from 46% 4 years ago. And with the close of Q1, we now have an RPO value that's greater than $99 billion. Now this RPO velocity and the accompanying revenue acceleration didn't just happen overnight. It's been steadily building as we've been going. It's included the gradual and ongoing conversion of our installed base of on-premise customers to the cloud. That's number one. Number two, it's about winning net new customer opportunities using the technology advantages than not only I just described, but that you've heard throughout the day this morning. And then third is our embrace of multi-cloud, which is giving us significant new ways to provide customer choice and driving more revenue opportunities for Oracle. It's these three things working in concert that's behind our bullishness. So I'm going to break down each of these three ideas into a little bit more detail for you today. So let's start with our installed base of support. We have approached or surpassed in many areas, feature parity across our cloud applications compared to our on-premise applications. Steve Miranda pointed out many of those areas in which our cloud is actually outperforming many of our traditional applications. The move of these customers has been accelerating. In fact, it's doubled over the last 4 years. Yet, as you can see up here on the chart, there's still a big installed base of application customers yet to move. Same thing on infrastructure, but we're even earlier in that cycle. In that case, similarly, we've more than doubled in the last 4 years, the movement of our infrastructure support base to the cloud. Now when our customers do move to the cloud, the amount of revenue we're generating is getting even better. So if we look at the last fiscal year, we experienced a 4x uplift in ARR in applications and a 5x uplift in infrastructure. That's even better than I talked about last year when we discussed these same multiples. Using these improved conversion rates resulted in a potential incremental revenue opportunity for Oracle of around $85 billion, just upgrading our installed base of customers. Now let's turn to the expansion with net new customers, where we're gaining interesting new workloads that we never had before when we were an on-premise company. So as you can see, the rate at which we're adding new customers is increasing. So it went up from 10,000 new customers in FY '23 to 11,000 last year. We now have over 80,000 customers that have bought Oracle Cloud. In terms of who we are attracting, there are four interesting types of customers that are being added. The first is we're winning more application customers, whether that's moving on-premise customers from our competitors like SAP or Infor or we're actually getting a lot of changes from existing cloud customers who want to move away from what I'll call legacy cloud like Workday. In addition, with the build-out of our mission-critical industry applications, we're just starting to see the uplift in the amount of growth we can see with landing customers in these areas in the businesses that Mike has been focused on like hospitality, construction and engineering, health care, industries of that type. We're very early in that adoption cycle. The second, of course, is AI. Two years ago, this category didn't even exist and now it's exploded as these companies are very well funded and have money to spend, but they're seeking the very best infrastructure technology on which to run. And Oracle has proven to be an excellent fit. Third is via the cloud infrastructure options only Oracle can deliver. Clay talked a little bit about this in his presentation, but things like alloy, where our partners can run their own OCI clouds in their data centers, nobody else has that. Sovereign clouds via governments and country-specific data centers, where with our flexibility in size, we can go into many, many more countries than our competitors can do to provide this option. And finally, dedicated where customers can take the same capabilities of our public cloud and put it into their personal data center and wall it off and yet we can still manage it and provide them the economics of public cloud. So given our flexibility, this group of customers is highly unique to us in terms of our ability to land them and grow our revenue. And fourth is ISV and native cloud customers, where they want to participate in the bigger Oracle ecosystem or where they have demanding needs where they see the price performance advantages of Oracle. So taken together, this whole group significantly expands the opportunity for Oracle to sell our cloud. Now the final area of momentum is being driven by multicloud, where, of course, as you've now heard, we make our software available via the public clouds of our competitors, who are now our partners driven principally by the Oracle Database. But before I get into the details of that, I just want to remind everyone how sticky the Oracle database is. It's not by chance that all three of the non-Oracle hyperscalers are now calling and are big partners with us. First off, the Oracle database remains highly ubiquitous. It's in use by 94% of the Fortune 100. Second, the Oracle Database is sticky. If you look at over the last 4 years, our net dollar retention rate in our support base is 102%. It's not declining, it's growing. And finally, the move of the Oracle Database workloads to the cloud is just beginning. So not only are these database workloads ripe to move to the cloud, but customers are showing an increasing interest in doing so. So they need to be on later versions of the Oracle database in order to take advantage of the cloud features that we have available, things like 23ai, which Juan talked about today. So we've seen a tremendous uptick, as you see in the slide, in the percentage of our database customers that are now on our latest releases. That's how we refer to customers in the database side as cloud-ready. The final thing that is helping with the move of these workloads to the cloud is customer choice. As we've now heard from various sources, almost every customer uses multiple clouds. The panel of customers up here help amplify that opportunity. And our goal with this is to provide the ability to use Oracle on the cloud of the customer's choice, while we work with each partner to ensure a seamless experience in using our technologies. So customers choose where to run, and we give them access to the very best features while also making it very easy to purchase and manage together. So now beyond running on Oracle OCI, we have partnerships with the rest of the large hyperscalers, Azure, Google and now AWS. Our work started with Google -- with -- sorry, apologies, with Microsoft and they really leaned into this early a year ago. And then, of course, this summer, we announced our partnership with Google Cloud, and it's amazing how much faster they are also ramping up their interest and the number of regions that are coming on quickly together. And then, of course, the trifecta was Amazon AWS, which we announced this month. Across each of these, we have very strong pipelines that are growing very significantly, and we expect it to be a more and more meaningful impact to our financials as we move forward. And in fact, we have over 450 joint customers with Microsoft Azure. You may have seen some of these stories. Obviously, you heard them from the panel, though if you walked around the show floor, you would have gotten many more of these stories. A lot of these customers have big ambitions. I can tell you that all of us as a management team are very pleased with what we've been seeing so far. So our final driver is how we're building this capacity to meet this customer demand and turn all of this momentum into accelerating revenue and profits. We've been on tear the last 8 years. As Clay mentioned, we really launched Gen 2 OCI in 2016. It was a very small footprint with very few services. And yet over the course of the next 4 years, from 2016 to 2020, we grew the number of megawatts under management by 20x. And then between 2020 and 2024, we grew the number of megawatts another 4x. The demand keeps coming and as you've all heard and continue to see, it's oftentimes faster than our ability to build out our capacity. It's a great problem to have, and we focus a lot of energy from an operational standpoint on getting that back into balance. Not only have we caught up with our major hyperscaler customer partners -- sorry, of competitor partners in terms of the number of features and services that we have on OCI but we actually believe we're in more places than any of them on a comparison basis. In fact, as everyone probably is aware now, Gartner now ranks Oracle as one of the leaders in the Magic Quadrant for hyperscalers. We've come a long way. Now the great differentiator for us, which I'll highlight again is that flexibility of deployment. The flexibility deployment allows us to do interesting things because we can -- as a result of being in a smaller footprint, we can tie up less of our capital in land and buildings and more of our capital in compute and storage. But to keep up with that demand, we've got to fulfill the backlog by building out the capacity that's required. So as you've now heard, we're going to double CapEx this fiscal year. That increase in CapEx spend is the precursor to accelerating revenue. So as you see, over the last 5 years, with more capacity coming online, the percentage of our software revenue that's coming from cloud compared to license and support has now doubled, and it's about 45%. We expect that percentage will continue to go up, given license and support is a slowly declining business compared to the hyper growth that we're seeing from the cloud. Now the kicker for us is the trade-off is actually very attractive from a total profit standpoint. When you look at the gross profit contribution standpoint, we are adding many more dollars to our bottom line compared to the decline from lost dollars to license and support. So while the cloud has more inherent costs in order to deliver the service, the additional profit dollars more than makes up given our growth rates. In fact, what we've been seeing is about for every -- we get about $5 of added gross profit in cloud compared to an average of $1 lost to the decline in license and support. So the bottom line is as we move to the cloud, we are able to accelerate the growth of both our top line total revenue as well as our bottom line total income dollars as we make the transition to the cloud, and cloud becomes a bigger and bigger percentage of our overall business. So let me just recap I think what you heard today. Our confidence in accelerating our revenue growth has only gotten stronger. First, it's because we think we have the best technology portfolio out there for what the market is seeking. Second, these product advantages are driving lots of customer momentum and interest that's ultimately driving bookings. And third, we're building the capacity to meet this demand. But what we are gaining is both more revenue and more profit dollars as we invest and grow our business in the cloud. Now what does this all mean for our financial outlook? What we've all been waiting for. Based on the strong backlog that we've discussed today, we are now even more confident in our ability to achieve our FY '26 revenue targets that we've previously issued. In fact, we are raising our revenue target from $65 billion to at least $66 billion for FY '26 while at the same time, we are focused on the bottom line and remain committed to growing EPS at least 10% in FY '26. However, given the strong demand that we've talked about and you've seen from Oracle, we've decided to defer an even higher EPS growth rate near term in order to grow our revenue faster over the next 5 years. So as the saying goes, just one more thing. As a result of what we've just discussed, we are sharing with you today our expectations for FY '29 to give you a sense of where this is all going. We expect that revenue in fiscal year 2029 will exceed $104 billion. This implies an average revenue growth rate of 16% between FY '26 and FY '29. While at the same time, we plan to accelerate profitability with the plan to reach the 45% operating margin and to grow EPS greater than 20% in FY 2029. Thank you.

Please welcome Larry Ellison to the stage.

Did he say $104 billion. That's going to be so easy. It is kind of crazy. Yes.

Thanks very much, Larry. Brad Zelnick, Deutsche Bank. Great to see you. Larry, you've been in the software industry for, I think, over 50 years. You've seen many paradigm shifts. You've figured out how to succeed, how to win across every single one. What is it today?

I'm actually having a very hard time hearing. I hope I'm not going deaf. It's probably -- I hope it's a sound system. I'm going to move...

Can you hear me now?

I can hear you now.

Cool.

Perfect.

So you've been in the software industry for over 50 years. You've thrived through multiple paradigm shifts. Ahead of us, we've got a major opportunity that Oracle and many others are racing fast and investing very much in. What is it that supports your confidence that this is good for the industry of software, and there may not be a redistribution of value, I don't know, maybe to semiconductors or elsewhere along the tech stack?

Well, actually, I'm not going to sit down again. Now I'll just stand up for questions. The really interesting thing about AI, and again, I used to give this speech a long time ago when I was a kid. If you really want to understand the computer industry, you really have to study the industry that it's most like which is the women's fashion industry. And you need to start reading W Magazine because they tell you what's hot and what's not. And the focus on AI, we're very good at the AI stuff. But you really can't train AI systems without data. A lot of companies are very excited about going ahead and exploiting artificial intelligence. And an awful lot of those models are being built at Oracle because we happen to have really interesting networking technology in our data center that allows you to build huge clusters by the way, that we first built for our database a long time ago. And then when we brought our database to the cloud, we built very unusual cloud data centers with these very fast RDMA networks. And that what made it relatively straightforward for us to build very fast GPU clusters using NVIDIA chips, where we can go to 32,000 -- huge clusters give you terrific performance because the problem is not just processing the data and training, it's also moving the data into the cluster, and we're good at that. But I want to go back to this idea of, yes, we're really good at AI and training AI systems. But we have this franchise where most of the world's important data is in an Oracle database. And I think the market -- for a long time, I've kind of chatted with you guys and sometimes you have confidence in things that we're doing and sometimes less confidence in things that we're doing. And I think maybe we go back 3 years ago, something like that. And I think you guys were pretty confident that we're doing well in SaaS. We've kind of proven ourselves with Fusion and then the acquisition of NetSuite, we were a solid SaaS player, and we were going to be a solid SaaS player. I think the jury was out -- I'll just say the jury was out for OCI for us becoming a cloud competitor. But what we have made our name on, the technology that we had pioneered, relational database technology, I think most people thought we were going to lose that franchise. The only thing that kept us from losing because we had Oracle only at OCI. We didn't have it at Google and Amazon and AWS. We had a huge amount on-premise. A lot of our customers didn't move those applications to the cloud. But I think there was a lot of skepticism whether we move all of that information to the cloud. And in fact, no one else invested in database during that period of time. What is the real -- is it Snowflake? Is it Mongo? What is it? Well, I think people -- database isn't really very interesting. If you check W Magazine, AI is hot, database is not, which is really interesting because -- I mean I think it's still called the information age. And unless you have your data properly organized, you can't use AI. It becomes utterly useless. It's an enormously powerful tool, and it has -- doesn't have the access -- really coherent access to your data. So you can't exploit it. So if you want to take advantage of AI, you have to do two things. You have to really do a good job of organizing your existing data and making it accessible. And then you have to have the appropriate AI tools, whether they're large language models or other sorts of neural networks, depending on whether you're doing -- some of the things we're doing -- two things we're doing in health care. One is which we're creating -- automatically using large language models to create doctors' notes, we listen to a consultation between a doctor and a patient, we create the notes, we use a large language model to do that. We look at a biopsy slide, very different neural -- we use a very different neural network to look at a biopsy slide and say, well, do we think there's cancer there? So those are very different. But you have to have the training data for all the biopsy slides, you have to have all the medical records accessible. The hospital does, the clinics do, the payers do, the insurance companies do, the National Health Service in U.K. does. They have to organize all of that data, all that health data coherently. So it's really a two-pronged problem. When you say, I really want to use AI, I want to take full advantage of AI, well, you can't do it unless you get your data in order for both training and inferencing. And there is no alternative to the Oracle database that I know of. I would love -- I think Microsoft SQL Server is not a bad product. Maybe it's the second best relational database out there since IBM got out of the relational -- IBM kind of got out of the business. Amazon, who's valued and I think they're an incredible company, AWS is an incredible service, but they really don't build databases. They take open source products and they put them in their cloud. They don't do a lot of database for R&D. They do almost none. Google has a big query, but it's really not a database. So why am I confident to think that the future is bright? I'm going to slightly rephrase your question. You're saying, why am I not worried that AI will suddenly go out of fashion and something -- the next hot thing will come up. And then we're kind of -- well, we're solely -- sole dependent on AI. And we've got nothing else supporting our, hopefully, a very big stock price, right, and that's supporting our growth, our EPS growth, our revenue growth and making our business better. Well, I think we have multiple ways to take advantage of the information age, take advantage of AI. Part of it is back -- doing a really good job with that database business, holding on to that franchise. And if you think about us holding on to that franchise, what's that worth in the era of cloud? It was worth quite a bit back when everything was on-premise, everyone built their own data centers. But in the AI world where data is an essential enabler, we have the best database, and it remains -- and it becomes more popular than ever. I think that's an important pillar to our future, and it's an important part of the value that we provide to customers. Our databases don't get hacked. They don't get hacked. Lots of really famous companies get hacked, we don'. It's highly -- our stuff is highly secure. We pay a lot of attention to security. We think we have huge differentiation on security and reliability. Our database, one of the unique things about Oracle is I can walk up -- I shouldn't, well, walk up with the gun and just start shooting our servers, but all the -- everything would keep running. We're the only database that can take multiple computers and high-speed network them together and create the illusion it's really one computer. We run one workload on 64 -- if you will, 64 servers with this high-speed RDMA network interconnecting them. And I can just start shooting 20 of them. And the system will keep running. No one else has anything like that. No one else has that kind of scale up. It's called RAC, it's called Real Application Clusters. We had it for a long time. You know who else has got it now? A decade or so later, 2 decades later, no one has it now, nobody. We're the only ones that could provide that kind of reliability. We're the only ones that we can scale -- I put up. Another thing in security, we're doing -- getting rid of passwords and we talked about what a wonderful biometric authentication and Google Pay, Apple Pay, which is kind of the beginning of biometric authentication for credit. But we can spread it to all -- any card, Mastercard, Visa, American Express, we can make it work with a biometric database all over the world and you need to be able to do 22,000 transactions per second to validate those trends, not a problem, not even close to a problem, but nobody else can do it. And it has one run 24 hours a day, 7 days a week, never not get hacked, never -- not a problem. Who else can say that? It's interesting. So we have highly differentiated technologies. Our networks are very different than our competitors. The RDMA networks that we build are essential for training large neural networks but they're also essential for building this extremely advanced, extremely reliable, highly scalable, these global databases that we can build. And it means we can build systems that don't have to come down. We patch our database while it's running. So if you're a phone company and you're expected to provide a service that never goes down. Well, Vodafone bought six of our data centers. And I think that's just the beginning because our database is also faster which is you say, well, yes, it's great. You go really fast. That's really nice. Okay, let me translate that to a slightly different synonym in the cloud, much cheaper, it's much cheaper. And that's why we have an advantage in training AI models because, yes, our networks are faster, which means training happens faster and it costs you less because you charge by the hour. And if we charge half -- if we're twice as fast, we charge half as much. So I think -- so let me just -- I think database is an unbelievable franchise and sustaining it into the cloud era is bigger than our business is now, everything -- more important than everything we do now. So I think the other thing, when it's fully grown, when that business is fully grown. I think AI is also an astonishing business. Last night, I had dinner with a friend of mine and my son at Nobu in Malibu, and we were talking about the future of robotics. And he's building a bunch of them in Palo Alto. So it's unbelievable what's going on. We'll be training these humanoid robots I was just discussing with Safra because Safra and I, we constantly say that every 10 sentence, I feel Like I'm living in a science fiction movie and then we kind of get over it and go back and talk a little more about business. But we'll be training robots to be nurses. So it's -- we'll be training robust to do a variety of things in hospital, but also nurses at home. These robots can do -- they were originally designed to work in factories and they're humanoid robots because they'll be doing jobs that were formally done by human beings. So you need a robot that's got two arms and two legs and kind of fits into the same spaces and could do the same manipulations that human beings can do. So they can work in restaurants, they can work in hospitals. And we're very involved in automating hospitals and there are a whole bunch of things we're going to do in terms of rolling out robotics in hospitals. We're connecting -- we're working -- oh, God, a big German engineering company that makes medical devices whose name I'm not sure I'm allowed to say, but it's really a big, important German company, begins with S, but I can't tell you any more than that. We think all of their diagnostic devices, there are two-tube scanners, their -- everything. Their x-ray machines, their sonogram ultrasound machines, all MRIs, obviously, all of that data, all of those machines have to be connected to the Internet. And then they're all part of our big -- our IoT framework. And we're working with these companies, whether it's bioMérieux in France or the S company in Germany or -- to connect all of these devices to the Internet. We have -- I think we're better at IoT than anybody because and a really interesting reason why we're so good about it, this is the only question I'm going to answer over the next hour, realize I'm going on and on here, maybe I should -- the question. Why are we good at IoT because we build applications that make demands on the underlying infrastructure, and we eventually get it through our heads exactly what we need to do because we're really doing the applied IoT. We actually have to hook up the medical devices in the hospital. We've got -- we're using RFID to track inventory in the hospital. We know how many RFID antennas and where they go and how much they cost, and we value engineer it because we actually deliver it. We don't -- I mean, I used to be a -- an anecdote real quick. I was a programmer working my way through college and onetime, one summer, I got a summer job. Anyway, they have -- I built an application, built an application program. Actually, it was automating an ice cream factory. And I don't know why they gave me that job, but they gave me the job. And I wrote the application. I thought I had done it perfectly. I turned it over to them kind of my last day, went back to college and they said they threw the application away because it didn't work. And I said, "Well, I don't think so. You guys make ice cream, I'm a really good programmer. I'm sure it works." And so I went in and say, what went wrong? Well, they were making chocolate ice cream but they had run out of chocolate. And I said there was no way to make chocolate -- when you don't have chocolate, you can't make chocolate ice cream. And plus you don't have enough -- and by the way, and you couldn't have made chocolate chip ice cream either here because you had -- you only had half as many chocolate chips you needed to make chocolate chip ice cream. So you say you made this stuff but I know my system shows it's impossible to make. So you're wrong, I'm not wrong. And they said, "No, Larry, if we run out of chocolate, we just use cocoa. And if we're missing half the chocolate chips, we just put in half as many." Why would you ever hire a college student to write that program? I had no idea. I didn't know. But now because we actually do it, we know, we built the IoT framework. We've got a bunch of things that were -- the kind way to say it was suboptimal. But as we built one application after another after another on top of that IoT framework, it got really good. It handled one use case and another use case and another use case. And there are lots of this -- so we had this unusual structure as a company that we are both a supplier of this foundational technology you guys call infrastructure. And then we were a user of that foundational technology as we build a variety of applications, build school safety systems, police automation systems, military systems, all sorts of applications on top of it. And we -- and we're able then, we learn, we're in that learning curve, we're able to learn and constantly improve the database, constantly improve our IoT frameworks, constantly improve our network performance and reliability, et cetera. And it's that tight loop, I think that gives us -- is one of our secret advantages versus our competitors that we're doing both. I'm sorry, my next answer will definitely be shorter. Over here.

Thanks, Larry. Jackson Ader at KeyBanc Capital Markets. If I think about 2029 what about artificial intelligence has to go right for those targets to be hit? And how right does it have to go for those targets to be hit?

Okay. I think on the AI -- I mean, our dependency on AI. By 2029, I can guarantee you, AI is not going to be the problem because the simple phrase, the race goes on. And I mean, this is like Formula 1. What do I mean by that? It's really not one winner. I mean you got three people on the podium, but it's really kind of one winner. Someone is going to be better in this than anybody else and multiple people are trying and there is a race. If you listened to Jensen Wong, and I'm sure you do, I know you do because I've seen the stock price. And I know -- I went to dinner with Elon -- in Nobu, Palo Alto, I went to dinner with Elon Musk, Jensen Wong and I went to dinner. And I would describe the dinner as Oracle and -- me and Elon begging Jensen for GPUs. Please take our money, please take our money. By the way, I got dinner. Please take our money. No, no, take more of it. You're not taking out enough that we need you to take more of our money, please. It went okay, it worked. I mean we -- yes, I mean, the demand for GPUs, the desire to be first, the desire to build the most capable neural network in the world, getting there first is a big deal, whether it's getting there first in self-driving or getting there first in reading cancer biopsy slides, getting there first in synthesizing video and make movies or I mean -- but there's -- I mean this impacts so many -- doing protein design, build design -- I mean we're very involved in designing both small molecules and protein, much larger molecules, peptides, proteins for cancer therapeutics, designing cancer therapeutics. I mean being first is very important. And the guys who are in this race are very smart and they understand they need to be best at something. They'd like to be first. So they're spending a lot of time and a lot of money, begging Jensen, building -- we're building data centers. I mean, my God, we're building nuclear reactors, you're kidding me, that sounds completely made up, but it's not. You need a lot of power to power acres of these -- I mean, it's acres of these GPU clusters. I'm not sure, has anything like this ever happened before? You know what, basic stakes is you know what it costs to build a frontier model. Anyone know over next 3 years, how much will be spent if you want -- you're one of the companies that want to build a frontier model, how much will you spend? Anybody? Anyone want to guess? $10 billion? Anyone -- $100 billion? Yes, $100 billion. Let's kind of gets you in the game, put down your $100 billion -- you ever see Molly's Game. You come in and you say, here's my money, it's not Molly's game. It's a bigger game. Put in your $100 billion and you're in the race. Not a lot of people, not a lot of companies, not a lot of countries will participate. I'm going to go over there. So that's good -- by the way, good for us. Very good for us. We're good to '29. We're okay.

Derrick Wood at TD Cowen. I think a lot of us view Amazon to be a bit of a foe to Oracle, at least in the database business, but they decided to partner with you. Can you just shed some thought on why they want to partner with you and then even stepping back from that, as you go into this multi-cloud capability, how do you feel about how that helps your ability to gain market share in the database space?

Yes. Well, I think, again, we've been a supplier to Amazon. Let's go back to the days before it was AWS. Amazon ran on the Oracle Database for a very, very, very long time and I think I got kind of cute commenting about Amazon uses Oracle, doesn't use AWS, blah, blah. And that hurt some people's feelings. I probably shouldn't have said it. But Amazon, we have a lot of the same customers. I think they're partnering with us because our customers have huge investments in Oracle databases, huge investments in applications built on those Oracle databases but they also like Amazon. So some of those customers want to move some of those applications to Amazon or some of them want to move them to Google, some of them want to move them to Microsoft, some of them want to move them to Oracle. I mean you've got all of the above. The only way you can do that or the best way -- I should say, the very best way to do that is get access to the very latest and greatest Oracle cloud technology. And we talked about this with Microsoft initially, and they said, "Yes, no, that's great. Customers -- it's what customers are asking for." I've had Amazon customers, a very big bank in New York and a very close, good friend of mine. I mean, he'd yell at me, but he said, pretty much, Larry, would you please make sure Oracle RAC and Oracle Autonomous runs at Amazon. Because we've got a big commitment to Amazon, and we're going to move a lot of our apps, not all of our apps, not everything. But I don't think anyone is going to move -- a giant money center bank is going to use -- move everything to one cloud, I don't think that's going to happen. It might, but I don't think so. They're going to use multiple clouds. And -- but we'd like to use Amazon. And I said, "Great. Makes sense to me." So we decided. Now it's an interesting problem just to do it is not easy to actually -- I mean, autonomous only runs -- only ran in the cloud. There was never -- it was never a software license you can get for the autonomous database. It was always a cloud service. It was born as a cloud service. So how do we take a cloud service out of Oracle and move it into Amazon and Microsoft and Google? It's not a matter as simply as shipping them some software and saying, "Okay, run it on your computers." That's not possible. So it turned out the only -- really the only way we could do it was to embed OCI into Amazon, Google and Microsoft, a miniature version of our -- not such -- a miniature version of our cloud, they could run all these different databases. In fact, run any of our -- all of our services because all of our clouds do the same. All of our clouds -- it's another huge difference between us and everybody else. All of our clouds are software identical. All of our clouds have all of our services. In that way, we have one automation system that runs all of our clouds. If we make all the clouds slightly different, we can't automate them. But anyway, step one, so step one was we got -- now we've got to shrink our data centers down small enough that we can -- because we can't build 50 huge data centers inside of Amazon on day one because they need them in several different locations, maybe even in several different countries. So we got to have some way to embed an Oracle data center in Amazon, Google and Microsoft that many of them and many do at different locations and do it in an economic way. And we made our -- our data centers are scalable, meaning you can have a starter data center today, which is about 150 kilowatts. And then you just keep adding more and more racks as you get more and more customers. But you can get the full Oracle cloud data center at 150 kilowatts. I mean this is -- I don't think any of our competitors are 20 -- that's 5% of -- and I'm sure 5% is a conservative number, meaning I don't think they can do it. They're 20x larger, their data centers are 20x larger than our smallest data center that we can put in. So we were able to actually then start with a bunch of small data centers in Microsoft. And as they have -- as they add database customers and they add Oracle services customers, we can make those data centers bigger and bigger and bigger. And we developed that strategy a long time ago because we want to put Oracle data centers in every major city in the world, in every -- pretty much every country in the world including -- we want to put Oracle data centers complete clouds -- by data centers meaning the complete cloud on a submarine or an aircraft carrier. We certainly can get in an aircraft carrier, no problem, submarine is a little trickier. The -- so that's another thing -- problem we had to solve to meet the customer requirement, I want to run Oracle Autonomous Database inside of Microsoft Azure. Well, then we have to embed the Oracle Cloud inside of Microsoft Azure. Then next question is, where is Microsoft Azure. Let me give you a map in these 60 locations. Oh, so you need 60 data centers, embedded -- Oracle data centers embedded in Microsoft Azure. In fact, the real number is we're building about 30. But who knows, we might be -- might go to 60. So we had to solve that problem, but we did. And we solved it in a way that has -- not only allows us to embed these data centers in Microsoft Azure and Google and the submarine and do things like that. But we did it with a high -- it allowed us to get a high degree of automation. And automation means you're much more reliable because if there's no human labor, there's no human error. It means you're a lot more secure because if there's no human labor, there's no human mischief or no mistakes that expose security vulnerabilities. And if there's no human labor, what is the cost per hour of no human labor. So our data centers are faster, more secure, more reliable but you got to be willing to pay less. It sounds impossible. But in fact, it was an essential to solve -- if we're going to solve the security problem, we have to have an autonomous database. In fact, we have to have autonomous data operating systems. We have to have autonomous data centers. They've got to be robotic. Robots drive cars way better than human beings. They fly airplanes way better than human beings, way more consistently, way more reliably and the cost way less, gives us a huge competitive advantage. So I think that's another interesting differentiator between us and the competition. We do applications and infrastructure. When you could say, well, Larry, you're not focused. You're not concentrated on infrastructure like your competitors. Actually, we think building applications has dramatically improved our infrastructure. We think the customer is demanding that we run in all the different clouds, turned out to be very good for us. And by the way, and for Amazon. So Amazon now can satisfy their customers who want to move their Oracle applications to Amazon. Customers happy. Amazon is happy. Our database, they keep using the Oracle database, which is good for us. So we're happy. I think it increases the size of the market for us dramatically and our customers have viewed this, it's been very interesting because we debated this a lot inside, okay, what's going to happen when we partner with Microsoft and Amazon and Google, we debate back and forth? Is this going to hurt our business? Maybe it'll help the database business, which hurt OCI. And how is this all going to play out? I think our experience and please ask Safra because Safra are the ones who spent a lot with Clay. We've spent a lot of time talking about this, of course, we're trying to predict the future, which is dangerous and complex. But I think the customer reaction has been so positive that Oracle is doing this, that customers are very open to saying, okay, yes, I'm using AWS, but I'd also like to use OCI. I think you guys are really good at this. And I'm going to have 2 cloud providers and the one's going to be Amazon, one's going to be OCI. So I think our OCI business has actually been strengthened. And you can make the other argument that it would hurt the OCI business, but I think as it turns out, the customer dynamics were such that it's actually helped our OCI business. And it's growing our database business and our OCI business is growing faster than otherwise would have. In the back there, gentlemen, yes.

Karl Keirstead at UBS. I've got a similar question but rather about Microsoft. It's been extraordinary watching the warming of this relationship between Oracle and Microsoft over the last year. That's interesting enough to me that I'd love to get your perspective. I don't think anybody is shocked that Microsoft would want a historical databases in Azure. You can see how they would benefit from that. What I think is extraordinary is that they'd be buying so much OCI capacity. That is wild. Can you give us some context?

Yes, I think -- we're very good at this. I mean, we're really good at this. And Microsoft is making a major play in AI. And I think it's more important if I'm running Microsoft which I'm not, but what is more -- I want a train, I'd like to see that ChatGPT. I've got an investment in ChatGPT -- openAI, ChatGPT. I have some of my own AI services that I want to provide. If Oracle's cloud can help me speed up the training of ChatGPT, it's good for me. Good for Oracle -- it's good for Oracle, but it's also good for Microsoft. I mean there's no -- there's really, in a way, no magic here. Other than at that point, you have to say, well, Oracle must be pretty good at this, AI training, that Microsoft would actually go ahead and buy it. And I think that's true. I think we are very good at it. But Microsoft is looking, I think, long term and broadly that they want to make their OI -- their AI -- excuse me, not OI, AI is as capable and as competitive as they can. And if one -- and by using Oracle Cloud services, helps them achieve that goal, so be it. Maybe they said, "Oh, I wish we could have done it ourselves. We didn't need Oracle." But if they can help us, let's not lose sight of the goal. The goal is for us to be a leader in AI. If they can help us be a leader in AI, let's go work with them. I think we're training -- it was -- I [indiscernible] with Elon last night and my son, and we're training Grok, Grok 2, and it's doing well. And yes, if I think same thing. I mean, Elon does a lot of things on his own because he's really a good engineer. I mean it's a great -- engineering team is great, brilliant guy, does that. But I think when you made the decision on get started quickly, start training -- start training Grok, we were the best choice, and he chose to work with us. And we're very proud of that, and we think that's a testimony to the quality of our offering at OCI, our AI offerings at OCI. And I think the same thing so if Elon picks us and Satya Nadella working with us, those are all good signs -- that we have technology that's valuable and differentiated. Okay. Right here.

Larry, it's John DiFucci from Guggenheim. So...

John, you wrote some very nice things.

Well, thank you because you drove that you and your team. Listen, I've really -- I think a lot of us here really appreciated some of your comments on the call around AI. We hear a lot of people talking about it and how they're going to charge extra and you, I think, use the word bewildering, which I think was accurate. Listen, I think we understand what you're talking about, and it makes sense that AI is just going to be part of everything. But there will be some things there already have been that have benefited from some areas of technology, whether it's OCI and the other hyperscalers or whether it's Jensen in GPUs. If you stay a step back from Oracle for a second, what other areas do you think are likely to be beneficiaries? And what other areas may be could see some harm or at least struggles because if they don't embrace it, some of the applications. It's easy to say, well, if you're an application vendor and you don't embrace it and include it in your technologies, you're going to fall behind. But can you comment on that?

The -- it's really interesting. We're -- we bought Cerner. This is another pillar for growth. I think you haven't quite seen it yet. But we bought Cerner and then we're in the process of rewriting all of Cerner. Now how can we possibly rewrite all of Cerner in 18 months. We -- maybe it will take us 24 months to fully rewrite everything they did. But I think that's about how long it will take, because we're not really rewriting it, we're generating it using our code generator. We're using code generators. And so what am I getting? But the user interface for this new medical system is very different. And I don't think I can charge separately for the user interface. For example, when you want to look at the latest equities, I took my son into the Stanford, he had a broken leg and they took some x-rays and the orthopedic surgeon couldn't find the x-rays at Stanford. They're using Epic. And then -- and my orthopedic surgeon is a computer programmer in addition to be an orthopedic surgeon. And he calls in the -- an x-rayed from Radiology, who's also a doc and she comes in and she's also a computer programmer. And she can't find the x-rays. And he took the x-rays about an hour ago. She can't find the x-rays with Epic. I mean it's -- and then they find me -- bring in the Epic expert at Stanford and she found the x-rays. Took 3 doctors to find my son's x-rays. This is how you find -- first, they have to log in and do all these other things and then find the right -- find -- go to radiology department, find diagnostic -- diagnostic devices. It was stored on one of the databases and one of the buildings at blah, blah, blah. They had to find all that stuff. This is how you do it with Oracle Health. Oracle. Show me Zeke Ellison's latest x-rays. That's it. Automatically logs in, looks at you, automatically logs in, finds the radiology database, brings -- knows you're a doctor who can actually look at Zeke Ellison's -- you're authorized to look at Zeke Ellison's. He's your patient. So you're authorized to actually look at that data and up pops the data second later. That's the difference. That's all and it's all AI. The -- we have a voice interfaces for all of this stuff. Another -- so how do I charge extra for that? No, I just think we'll sell more Cerner when we're competing with Epic -- that's what I think will happen. Another thing we do, and I probably have told the stories too many times that a doctor is meeting with a patient. One thing we do is very interesting. When the doctor schedules their day, we mix in-person meetings with telemedicine meetings. We make no distinction for telemedicine. So telemedicine is just built into the system. So it could be a remote kind of -- we use our own Zoom like telemedicine, HIPPA-compliant telemedicine secured conferencing. But before you meet with the patient, either in person or via telemedicine, we prepare a summary. We use a large language model to prepare a summary. It goes in and looks at your latest laboratory results. It looks at the latest entries in your electronic health records, It looks at what medicines you're currently on and prepares a summary for the doctor, that they can read very, very quickly before they see the patient. The patient comes in and the doctor consults with the patient. We listen to the consultation. This is all AI. We listened to the consultation. We then prepare because we're not board-certified and these, we can't actually update the electronic health record directly. We have to prepare a draft for the doctor of the updates to the electronic health record, to the new prescriptions, to the orders on discharging the patient from the hospital, doubling the dosage of lisinopril from 5 milligrams -- from 2.5 milligrams to 5 milligrams. So we collect all the -- send the prescriptions to the pharmacy. We send the orders to the nurse's station. We automatically update the electronic health records the moment the doctor reads the summary and okays it. And all of that is AI. We don't charge separately for any of it. It just how the application works. The application just works much better. The doctor is spending -- not spending any time in front of the keyboard. So it makes the doctor much more efficient. It has more value and people are more likely to buy that than something where they just take 3 docs in a room to find some x-rays. And I think you'll see that throughout this -- our system. We use -- the autonomous database, we use AI to build the autonomous database. AI is the autonomous database, which is built on AI. It's completely self-driving database, robot DBAs. Again, there's nothing like this. Just like there's nothing like the RAC thing where we connect a lot of different computers together and create the illusion as 1 computer. No one's done that. No one's done an autonomous database with no DBA. There's no configuration. There's no -- that's all done -- that's all AI. We don't charge more for the autonomous database because it's autonomous. We think it's, okay, it's more secure. You use less labor, it's more reliable. All of those things, but -- it's just a better database. Go ahead, John, you have a follow-up.

It's going to be part of everything and everything is going to be better. But there are some incremental things like NVIDIA and OCI. Any other like...

Well, we'll sell more application. We think if our applications take advantage of AI and our database takes advantage of AI, we'll sell more of it. The separation is completely arbitrary, but no, but I can -- I mean, we could have a credit card fraud detection or expense report fraud detection program that we see, okay. Fusion accounting will bill you extra for our expense reporting system, the fraud detection module is a separately recharged new AI module. We could do that. But in a way, I think it kind of misses the point. Yes, we can always take features, separate features of any kind, whether they're AI or not, and we think if they're a particular value or another way, very few people want them, not -- if everyone doesn't want them, you can't raise the price of the product or you don't really need to -- we're not trying to raise prices. We're trying to increase the volume of sales. And I think that's a better -- that's more our strategies than kind of looking at AI. Now how do I get -- I'm investing a lot of money in AI, how do I get return on that investment? Well, we want to sell more Fusion with Large. We want to sell more autonomous database with Large. We want to see more people to come to OCI. And I think -- and it would be an exception for us to say, "Oh, that new particular AI feature we want to charge separately for." Yes, sir. All right, right in front of me -- finally.

You've been great at setting long-term visions and then delivering on it. Oracle has taken a different approach to security in the cloud, delivering security out of the box versus everyone else, IT has to turn it on or not, rather or decide what to turn on or not. Now you're embedding more security with Zipper. You're building almost a security platform. I'm not sure exactly what you had even defined it as at this point of a security layer. Do you see this as something that adds value to just the Oracle products? So do you see this as something new innovative that becomes almost a platform or a product or a technology all into itself because security is such a problem and it's not being solved very well.

By the way, I could not agree with you more. But I think this is the perfect -- let me -- so a long time ago, Oracle had 2 versions of its database because our first customer was the Central Intelligence Agency. Oracle's very first database was -- we had a lot of security features in it. And we had these 2 versions of the Oracle database. We had one with -- and then -- and we helped the right this thing called the Orange Book. Now there's probably not the person left in the room that even knows what that is anymore. The -- but I'm old enough that unfortunately, I know. We help to write the security book. And we had these 2 versions. I remember being in a meeting and it turns out it costs money to have 2 versions of anything. We have to pay -- we have to build a secure version, and then we have to build a slightly different version without the advanced security features. And I remember be in a meeting and said, "Okay, who wants the database that's not secure?" What? It's a little bit like -- you've heard the joke the [indiscernible] comes into the rest of -- be coming to restaurant and what are you drinking? Water, water, water. And 1 guy says, yes, I want it in a clean glass and the waitress or waiter comes back and says, giving -- putting the water, "Who had to clean glass? Who asked for the clean glass?" The same -- what the hell? Who wants the not secure version of the database? What are you? Nuts? I mean, it's costing us extra money to build a not secure version of the database. Are we out of our minds. So I mean I'm not quite answering your question, but I'm coming close, and I will get there. So we -- you can't turn it. Actually, I had this argument with our Head of Database, his name is Juan Loaiza, who I love, he's a brilliant guy. And it's a long time ago, and I said, okay, Juan, from now on, everything all our backups are encrypted. There you get the choice, everything is encrypted, you can not -- but Larry, encryption costs -- it takes time, it costs more money to encrypt it. We got to use a computer to encrypt it, to decrypt it, blah, blah, blah. I don't care. I don't want to be the guy that loses your data. No, you don't have -- you don't get the choice eventually, took me a while, but eventually, I won that argument. And the -- so right now, you don't turn on. I'm going to encrypt this in an Oracle database. I don't think so. No, everything is encrypted, everything. You don't get a choice. Data in transit -- in storage, it's always encrypted. We're very careful about all of the stuff. And we don't give you -- when you do recovery, we don't let you do recovery. You just say recovered to this point in time or we pretty much say recovered at this point in time, then hands off. We don't let do that. We don't let people fly these things. What if you make a mistake,in recovery and lose your data. Oh my god, you're going to blame me. I know I'm going to get the call. No, I don't want that call. So we've taken away a lot of the choices. That's kind of why we can't charge it. It's kind of related to what John was saying, we can't charge separately for security. A lot of people do, that's a real -- it's a great revenue opportunity. You could have the unsecured version. But the guys with a little bit more money, they get the secure version or someone who just forgets to buy the secure version. We didn't know the secure vision was available. I mean, because a lot of this stuff is so complicated. We decided that security -- I don't want to read an article about us getting hacked, losing data. One of our customers -- hospitals being ransomwared. No, I don't want to take that phone call. I don't think Safra wants to take that phone call. So we are just making our system more and more secure as we have customers, not our systems that are failing. We have customers that certainly have been hacked and ransomwared, and a big part of the reimbursement system for the U.S. Medical -- for U.S. medicine was down for a while. I'm not going to start naming names, but thank God, it was not us, and we were able to help a lot of the customers recover. But we think security is another example -- it's always on. It's a part of the system. It's not an option. If we can make it more secure, we're going to make it more secure. We're going to make it absolutely reliable. You can't make a mistake. You can't, oh, I want -- no, you ask for a level of the reliability. You can say I want if this data center is hit by a media and is destroyed, I want to make sure I can immediately sail over to another data center. Yes. And then we have to -- if you want to do it instantaneously rather than just recovering, let's say, in a minute, you want to recover in a 60th of a second in electrical cycle, we can do that, but then that actually costs more because you have to have a full running system in another location. So you may have to, in that sense, pay for the redundancy, for the reliability. But we don't think it should be -- we don't think those are revenue -- maybe we're foolish, but we don't think those are good ways to make more money. We think, provide security, high degrees of reliability and then just sell more of it, to get more customers. On the back

Marshall Bush, Baleen Research. This is always one of the most fun hours of the year. So if you think that AI is going to drive share shift at the application layer. Does that mean that the ultimate value capture will be at the infrastructure layer because AI-driven applications will inherently be more chatty with the database in the middleware?

Yes, really interesting question. I think it's -- I think the applications that really do a good job of exploiting AI are going to do very well. And I think the race to build the best training systems, the race to build the best large language models, the race to build the best robotic models -- the robotic models are very different. The LLM is interesting. LLM is very different than because it's -- ChatGPT obviously, began this revolution in the sense that, "Oh my God, it talks, baby talks." And it was completely unexpected. But it's not real time. In other words, once -- you have to prompt it sometimes in a conversation. Think about how different it is when you're a robot. Let's talk about a 4-wheel drive robot called this Tesla self-driving car. And you see something in the environment and you have to react in a tiny fraction of a second. It is real time. It's a real-time neural network, has got to be. You see a ball roll out in front of our -- the car coming off the curve, and you see that, what do you do? You have to make a decision instantly as to what to do. That's very different than carrying on a conversation. It's very different than looking at a cancer biopsy slide and trying to detect cancer in the biopsy slide. Those are all somewhat different problems. They're going to be solved. They're going to need specifically trained models that are somewhat different, and they're going to need the appropriate applications that are quite different, one from another. So I think there's -- so in a way, I don't know how to answer the question. I think there's opportunity -- significant opportunity at the infrastructure -- at the AI infrastructure level, and there's significant opportunity at the application level. I think this is a really big deal, and it's at every place. Over here.

Brent Thill with Jefferies. With the success of OCI, do you think the application is great will meaningfully accelerate? I know that you mentioned the backlog is driven by OCI today, but is apps going to catch up and be the next leg here?

Yes. Well, I think Cerner would -- now called Oracle Health. It's the biggest industry in the world, a multitrillion dollar industry that hasn't been served by the IT industry at all. The 2 giants were Epic and Cerner. We're used to competing with Microsoft, Amazon, Google. Epic and Cerner don't have those kind of resources. They don't have that level of talent. So I think our medical system and we -- by the way, we're trying to do -- we're trying to automate the entire medical ecosystem. What Cerner and Epic did is they automated acute care hospitals. And what Epic specializes in is in the United States, Epic is the #1 company for automating academic research hospitals. That's what they are. They hardly exist outside the United States. Like we're #1 in Germany -- I mean, we're almost the entire National Health Service in the U.K., Oracle Health now or Cerner. So that's it. I mean, yes, there's Allscripts. There are a couple of little things, but there's -- how could the medical industry be completely underserved in terms of application software. It's stunning. How awful those systems are and we actually -- and actually, the purchase of Cerner has had an incredible impact on Oracle in terms of fundamental cultural change. I remember I made the statement of are actually able to fix this problem, actually do a good job of automating health care around the world. If we can do it, we must do it. We have a moral obligation to do it. I can't believe that this industry that is so important and touches all of us and all of our families, doesn't have the very best technology support by our industry, but it's been largely ignored. And people have doubled it and tried. And to really change the medical industry, you have to look at the entire ecosystem, not just acute care hospitals. You have to look at ambulatory clinics. You have to look at community hospitals. You have to look at medical laboratories and diagnostic laboratories. You have to look at payers, whether they're insurance companies or national governments. You have to automate because the negotiation, the preauthorization of -- you go into a hospital and you're having a hip replacement, the hospital can't do the hip replacement without talking to the payer, whether that's the National Health Service in the U.K. or it's an insurance company in the U.S. and get preauthorized, yes, if you do the hip replacement, you'll get paid for it. And there is the human beings on the phone discussing this, "Are you eligible." This is all crazy. This is all needs to be automated. The preauthorization needs to be automated. We need to -- in the NHS in the U.K., there -- these waiting lists. You might -- you need to see a doctor, and you can't see a doctor for 6 weeks for something. Well, you're calling 1 clinic, what if there's a doctor available in another clinic 8 miles further away we can send you to that has an opening that can see you tomorrow or the next day. They have no idea. They can't do that. They can't manage the load, the request load and standard proper queuing theory, we'd call it, and they can't -- they don't have systems to do that. And we're managing everything from people on smartphones. That's how you make an appointment. So that's how -- we give you the -- you can see -- you go to your regular clinic. If we can get you in on time, we'll then give you the option. If you don't want to -- your regular clinic and see you in 4 days, but we can get you in tomorrow at a clinic that's a little bit further away. Do you want that? Do you have to have a patient engagement system on your smartphone to make the appointments. If the doctor is running late, we don't want you to come sit around in the waiting room for a couple of hours. We'll let you know in advance that, okay, we're running 2 hours late, please come a little bit later. All of this stuff to make things more efficient the doctors shouldn't be typing in on keyboards, that all should be done with AI, redoing all of that. How big is that? Well, that alone is much bigger than our apps business is now. We're talking about all the hospitals in Germany, all the hospitals in France, all the hospitals in Australia, all the clinics in Germany, all the clinics in France, all the community hospitals, all the pharmacies, all of diagnostic laboratories, plus the governments who are usually the payers need systems to do all of that. Plus we have need -- COVID made it very clear we need a system for a national public health. We have no system for national public health. We had no idea. We set a ship to New York City well intentioned because we thought New York City is running out of hospital beds during COVID. No, they weren't. Almost no one went on that ship, but we just didn't know. We didn't know sending people back in the nursing homes was killing people. We took a long time. We didn't know COVID had spread months before from Wuhan to Milan and then to New York. We caught it months after we should have known. And it's very easy to know when an epidemic or pandemic is beginning. We can -- you can go back and open source satellite pictures, open source satellite pictures and look at the parking lots in Wuhan and see they're filled at 4 a.m. in acute care clinics, something is going on. We have no early warning systems. We have no global systems for -- pathology systems that look at -- that actually gene sequence, a new coronavirus and say, well, that's a little bit different than I've ever seen before. We've done that. With the University of Oxford, we built a global pathogen system. We -- we've made -- we've attached gene sequencers toward the Internet of Things. So any hospital can quickly sequence a suspicious new pathogen, if you think it is something you haven't seen before. Yes, we -- I mean, of course, everyone does PCR and all these other things. All the PCR devices are also attached. The new PCR devices are also attached. We're doing all of that. So we're not just doing -- we're not doing what Cerner did. We're doing what the Cerner did plus a lot of other things to completely reform and automate and digitize the health care infrastructure for the world. And we have to do it in a way that it's pretty economical that we can do that, that we can do community hospital. How do you even communicate with community hospitals in Rwanda or -- that's East Africa or community hospitals in places in Central America or New Guinea. There was a gentleman that asked me a question a couple of days ago from New Guinea. How do we get the latest greatest digital technology in our country? We're not rich like some of the other people here. Well, we can communicate with all the community hospitals using Starlink. We can use modern satellite technology to communicate with any school, any hospital, collect all of this data. So we have a true early warning system for pandemics. The buyers and the providers -- see, I should say, the payers, the governments and insurance companies and their providers have a very easy way to get a prior authorization that's highly automated, doesn't make human doctors getting on the phone, begging for the latest cancer drug. We can use AI to read a insurance policy, we can use AI to help you understand your legislation and to see whether that's covered or not in most cases. You can look at the patient and the condition of the patient and tell you right away whether that should be covered. We can do all of that. And that's what we're doing. And that's one example. That's 1 example. I'll give you a couple of other ones real quickly. Securing schools. We think we can absolutely lock down schools so that dramatically reduce the case of anyone being on campus that doesn't belong on campus and immediately alert, second someone pulls out a gun, immediately alert -- use AI cameras to immediately recognize that. But, a, we keep everyone off-campus that doesn't belong on campus. The police another thing, body cameras. We've completely redesigned body cameras. Our body cameras cost $70; normal body camera cost, I don't know, $7,000. Our body cameras are simply lenses, 2 lenses attached to your best, attached the smartphone that you're wearing. And we actually take the video that the police officer is -- by the way, and the camera is always on. You don't turn it on and off. And by the way, the way you turn it on, you can't turn it on, I'm going to the bathroom. Oracle, I need 2 minutes to take a bathroom break, and we'll turn it off. The truth is we don't really turn it off. What we do is we record it so no one can see it, but no one can get into that recording without a court order. So you get the privacy you requested. But court order, if you get a court order, we will judge -- can order -- I want to look at that, this so-called bathroom break. Something comes up, -- and -- but I'm going to lunch with my friends or going in an hour for privacy with lunch of my friends. God bless, we won't listen in, unless there's a court order, but it's interest -- and - but we transmit the video back to headquarters. So headquarters and AI is constantly monitoring the video. Remember, this terrible case in Memphis where the 5 police officers basically beat to death another citizen in Memphis. Well, that can't happen because it would be on TV at headquarters. Everyone would see it. Your body cams will be transmitting that. The police will be on their best behavior because we're constantly recording -- watching and recording everything that's going on. Citizens will be on their best behavior because we're constantly recording and reporting everything that's going on. And it's unimpeachable. The cars have cameras on them. I think we have a squad car here in some place. But those kind of applications using AI -- if we can use -- and we're using AI to monitor the video. So if that altercation it occurred, they've occurred in Memphis, the Chief of Police would be immediately notified. There's not people that are looking at those cameras. It's AI that's looking at the camera. No, no, no. You can't do this. It would be like a shooting. That's going to be immediately -- that's going to be an event that's immediately an alarm is going to go off, it's going to be -- and we're going to have supervision and there it every police officer is going to be supervised at all times. And the supervision will -- and if there's a problem, AI will report their problem and report it to the appropriate person, whether it's the sheriff or the Chief or whomever we need to take control of the situation. We have -- same thing we have drones. We -- if there's something going on in a shopping site, and I'll stop, a drone goes out there and go there way faster than a police car. There's no reason for, by the way, high-speed chases, you shouldn't have high speed cases between cars. You just have a drone follow the car. I mean, it's very, very simple. I mean, new generation of autonomous drones. Spotting forest fires, a drone spots the forest fire and the drone then drops down and looks around to see if there's a human being near that heat bloom and someone else either had an unattended camp fire, that caught fire or it's arson. We can detect all of that. That's all done autonomously with AI. All these are AI applications. And so the -- if you think about the application business, and I'm just getting started. I mean I'm sure you have -- you have other things to do in your life and just listen to me. But there are so many opportunities to exploit AI. This would be the last one. We use satellite images to look. We can find all the farms in Kenya, all the farms in Rwanda, all the farms in Morocco. And we can look from the satellite, we can tell what they're growing. They're growing maize. We can report back if the northern part of the field needs more nitrogen, more fertilizer; the southern part of the field needs more water. We can predict the -- we're using AI, we can not only identify what the crop that's being grown, we can forecast the output, the agricultural output of that country based on what we're looking at now. We can actually survey and forecast the, if there's going to be a shortfall. If there is a drought that's going to reduce the output, we can provide early warning to the people in the agricultural department of that nation state. We provide all that information. So these are the kind of systems we can -- next-generation systems, we can build using AI. So when I say back to the previous question, yes, I mean, it's going to enable so many -- so much innovation in so many different areas that the world is going to be a better place as we exploit these opportunities and take advantage of this great technology. Safra, am I out of time? Okay. Thank you all.

Please welcome Safra Catz to the stage.

Thank you all.

Thanks, Larry. Okay. I don't know if you have any questions left, and I know a lot of you have flights. So we'll make this quick. Who's got a question? Can you?

A lot of excitement. It's palpable walking the show floor, talking to customers and partners, it seems like it hasn't been this exciting to be at Oracle for some time. I wanted to ask you about the fantastic guidance that you've provided today and that Doug went over with us. Can you maybe just talk a little bit about -- obviously, you're not going to box yourself into a corner and tell us line by line, exactly how you get there. There are many ways I'm sure that you can win. But as we think about the leverage in the business, both gross margin and OpEx efficiency on the way to '29. Can you just talk a little bit more about the multiple ways to win here?

Yes. So we have a number of ways to get there. And the reality is, first of all, you can see already in our remaining performance obligations. We have enormous amount of business coming our way. And this is only the beginning. So there's the OCI layer, the GPU business is part of it. The other part of it, of course, is the database business. And what's wonderful about the database business, especially in our multi-cloud approach is that, that is a high-margin business because it's an extremely high-value business. And if you've been listening to my calls and poor you, you have been, you know that we talked about 3 legs to the stool. Fusion, NetSuite and our vertical applications, it's already a big number. Larry just told you, Cerner, as you know, as I was breaking down over our first 2 years, Cerner is actually a drag currently. It's going to turn around and be like a neutral because -- and then it's going to be truly additive. And when I say neutral, I only mean that it won't be shrinking, but it may not be growing as fast, at least the first year as Fusion and NetSuite and then all of our other industry applications, they are only now coming online into the cloud. So those numbers are going to be very helpful. So there's OCI and the GPU business which the GPU business is only a subset of OCI generally, and then the applications and now the database. And sort of everything is hitting and the database has just started even though like for other people, it would be a complete company. But for us, compared to everything else, it is just moving to the cloud. And we've given ourselves a couple of years and basically everything is happening, and there are a bunch of other things that I'm not going to share with you now that we are just starting, that are just now you'll start to see little hints of it over this next year. The lines of businesses that you might have heard us talk about transactions and things like that, other capabilities that we've not even hinted at, all of those are going to start showing up over the next few years. And of course, we're not going to leave our historic discipline. We treat this money like it's our money, not necessarily your money. We treat it like it's actually our money. And so we're always extremely careful about matching up what we do without giving up profitability. And the thing that's most interesting is not even over $100 billion number, but the 20% EPS growth. I remember last time, we talked about that, and we delivered it. And then we did our transition. Every single thing we said would happen over these years, has happened. And do not think that we do not remember those of you who've been with us this whole time who have had an understanding and faith in our ability to ultimately execute this. And the success has been absolutely a result of some of the things that Larry shared in his original keynote, that he shared here that the technical teams shared. We have extremely differentiated products. We're not a me-too operation like, "Oh, they're doing this, us too. No. That's not our way. Always do the higher-value things, do it faster, cheaper and more securely. Those numbers should not be a problem. You pick.

As part of this new framework, how do I think about CapEx spending? And one of the things that could be interesting is like, if you sit in Azure, if you sit in GCP and AWS, they want to spend a lot of money for you, but like how do we have to think more broadly from broader picture perspective about the supporting numbers on cash and CapEx?

Well, See, you don't -- I don't want to say this so I'll just not. But I worry about CapEx, okay? And I figure out how to maximize the power of every CapEx dollar. So we are very, very careful to try to leverage every dollar and to also ride the coattails of richer companies who want to give us floor space and power. That's 0 CapEx for me, okay? And all I do is put in my computers. And when they've got floor space and power, my contribution CapEx-wise is very, very small. Additionally, there are all sorts of other clever ways to leverage other people's spending on your behalf and to make sure that we've got exactly a nicely aligned spending model. We will be spending more, there's no question. But we're always looking at clever ways to make our cash go further. And sometimes, as I've told you on different earnings calls, sometimes the difference between it in 1 quarter and another, you should not break your heads on that. You need to kind of look at it in a backwards 12 months or forward 12-month because within any 1 quarter, and I think I gave an example even on the earnings call, though, I feel like I've been talking so long. I can't tell what I -- where I set it. But sometimes, the components come to us. And when we buy them, they are capital spent. But sometimes, they actually don't come to us. They go to our one of our manufacturers, and I don't buy it until another 3 weeks later, in a computer. And you'd be like, "Oh, wow, look at that $400 million. Why didn't you spend it this?" Because I'm spending it 10 days from now. So don't worry about that. I've told you we're going to double our CapEx. Believe me. And of course, you'll see the revenue. The revenue is right there. And so double this year, not giving guidance on next year yet, but we'll get there soon enough.

Siti Panigrahi from Mizuho. The topic this week is all about multi-cloud, and it's very impressive to see the 3 partners in just 1 year. So we know that, that's going to unlock a lot of value for your legacy customer, database customer. So how should we think about the contribution of that these partnership going forward?

Well, once they get really rolling, it's going to be, I think, quite significant, if you want to know the truth. And it's going to be very, very profitable for us. I want to make one other point, though. You should understand that in this kind of multi-cloud world, it's not only them. It's NRI that I think was on stage here just a little while ago. It's Saudi Telecom. All of these alloy partners also have our cloud completely embedded on their floor space fully plugged in. But you know where else it is? At a bank, they have the full cloud. When they have what we call a dedicated region, cloud customer, they have -- it is literally a region. It just sits on their floor or whatever floor they told us to go to, and it's just theirs. So you need to understand that as -- Amazon is going to be wide open. Azure is already in action, Google went live during the show. This is a big part of it. But you also -- if you're hearing me on the calls, you hear me talk about consumption revenues going way up. Well, remember, we have planted in that big map, but we've also planted at customer sites. And as those fill up, that's money we spent a while ago. The JWCC, for example, many of you don't know what that is, remember, there was a supposedly JEDI project for the Department of Defense. Remember, it was canceled. It was a single source to one of these guys. It was canceled. Instead, it was the JWCC. We are the ones that got the single largest task order about 1.5 weeks ago. Remember, you have to understand, that data center is fully built out to federal standards, accredited, certified and everything. It's been ready for a while. We got the single largest task order, okay? So you -- so this is just happening. There's a lot out there. There's a lot going on, honestly. It is -- and everywhere we go, I want you to remember 1 thing, everything everywhere. So at Azure, it's all our services. They're available. They may not have been turned on. They look exactly the same. And if customers want them turned on, at Azure, at Google and ultimately at AWS. The folks at Microsoft have been rolling out services on OCI this whole time. They're actually all fabulous.

Okay. We'll take 1 last question as I know people have planes to make, and we're wanting to keep on time here.

It's John DiFucci from Guggenheim. So I want to come back to the database migration to the cloud because this is something I've been writing about for years, and it's been hard because it just makes total sense, but the timing has been tough. And it makes sense, Larry, you talked about these being mission-critical workloads are really important. So it makes sense that they are the sort of the last things to move to the cloud or when you're confident about the cloud. I've never lost faith in this like Brad and Raimo and Karl. But I -- but what gives you the confidence today that now it's about to happen? Is it -- do you actually see it in CRPO and RPO? Are those they contracted for now as part of that? Like do you see that momentum in addition to -- if you can just talk a little bit about the why now?

And I actually had it on stage with me that even though my customers, they all did different things when you're BNP Paribas and you take all your data, your Oracle databases and you put them in our cloud, it's starting. German banks, more and more, it is just -- it's scary. It's your company's heart and you're thinking yourself, "Oh, my gosh, is this one -- am I really ready to put it in the cloud?" Now it's happening. In fact, it's interesting because our multi-cloud strategy, which is not only being in the other hyperscalers, but allowing these customers to have their own region where they don't have to share with anybody. It feels to them like the safety of on-premise with all modern capabilities. They can go downstairs 4 levels down into their basement and be like, it's here and yet it is in the cloud, and they have all the services they want, they are completely up to date. By the way, one of you asked a question of Larry -- you did, I think, who are the guys that aren't going to survive this? I'll tell you who, folks who say their cloud, but are actually just hosted, okay? Applications where you get a new update every 3 years instead of every 90 days, those folks that game -- it's like game on now in applications. All these new modules, all these new agents if you're pretending to be cloud by just putting yourself in one of these hyperscalers buildings, that's not actually cloud. Cloud is new capabilities, exactly the same code, everywhere and new abilities and capabilities every 90 days. And so this is going to be a very, very exciting time for us at Oracle. You can tell -- I must say the funniest thing I heard, and it kind of goes against our whole messaging was one of our customers stood up at an event like in this room, and she said, or he said, "We're not multi-cloud. We're only OCI, and we're thrilled." So -- and we embrace that, but we also embrace all the choice. And this is a really exciting time to be with us. I have to say one of my customers in our public sector -- in a public sector meeting stood up at the city of Atlanta. They went Big Bang Oracle Fusion, Big Bang Oracle Fusion all the applications, and they just couldn't wait to talk about it. So with all our other customers. I hope that you all have gotten to see and hear from our customers. You shouldn't have to believe us. Of course, please believe our financial statements not only announced on day 9, but the Q was filed overnight. I want to thank you all. We do really appreciate your support. We have been working hard on this. I can tell you what Larry and I are talking about. We're talking about how exciting the next 10 years are going to be at Oracle. So you should keep an eye on this. We're having a lot of fun. Thank you very much.

Okay. I certainly can understand if you all feel that it was a little anticlimactic after the safe harbor slide. So -- but I do appreciate you all sticking around for this. Covered a lot of ground. As always, thank you very much for coming out. Safe travels home. Good night. We'll call it a wrap.

That concludes our program. Thank you for joining us and enjoy your day.