Informatica Inc. (INFA) Earnings Call Transcript & Summary

Well, this is always a fun annual event for us now. So like every year, I'll do a summary of the year and you tell me how you think about it. So since we spoke in March 2024 now, you've now achieved even more $1.73 billion in ARR, an $800 million cloud subscription business growing over 30%, cloud NRRs of 124% plus, close to 33% operating margins, and you pretty much remain the undisputed leader in most of your Gartner quadrants. And you made some good progress on CLAIRE AI, which we can talk about and a lot of new partnerships.

So can you tell us the journey of this last year, what you've learned, what you expected and what maybe we didn't go up to expectations?

Sure. Thanks so much, Sean. Yes, 1 year has gone by. A quick recap of the year, and then, of course, we'll get into what we think were -- what worked and probably what didn't work for us or opportunities for us. Look, for many in the room who may or may not know us, Informatica, we just clocked 30 years of our existence. And it's one of the rare companies where, obviously, we were a leader in what we did in a very small market as a small single product company. And fast forward now in our 30th year, obviously, we have reinvented ourselves, but from a single product company, we are a multiproduct cloud platform company, so on and so forth, a huge, huge transformation of a company that was public, went private and then obviously we went public again. But if I look at last year in particular, last year was a very important year for us because in 2023, we entered the last chapter of our journey, which was we sell only cloud. And in that next -- last year was the second year of that, and if I look at the 3 pillars in which we look at performance, one was innovation. Ultimately, in any tech world, if you don't innovate, you die. And our belief has been that we build the best products, and we put hundreds of millions of dollars of R&D in that. And it's as evidenced by each of the magic quadrants that you can take a look at. None of the start-ups, nobody ever has able to get ahead of us on any axis, both vision and execution. And it's not just that, whether it's the Forrester Waves or the IDC, whatever you want to look at, innovation has been #1. And in that best products, but we are uniquely positioned as the only data management company with a full platform where all of these products sit on a single platform, consumption-based pricing because the landscape around us is so fragmented. There are like hundreds and hundreds of point providers in each market we compete in. And that platform IDMC clocking 110 trillion -- 120 trillion transactions a month, growing 20%, 30% now is a very healthy place to end the year. Second is in terms of what I call customer stats, the business stats you talked about. Look, we ended the year very importantly, cloud-only $827 million of cloud ARR growing 34%. And we've guided this year to $1 billion growing at 25% and we'll talk about that. And of course, that's been a great business growing with an NRR of 124% in Q4. We got 5,000-plus total customers, 2,400-some cloud customers, and we feel very good about how they're growing with multi-use cases, and we'll go into how they are. And that's been a very sticky and a very healthy business to grow at. And what I feel very good about the fact is that if you take these 2, while we're doing an on-prem to cloud transition, we've navigated that with maintaining healthy gross margins, as well as high operating margins you just talked about and managing this from 2 by balancing the 2 without kind of going into the negative territory on overall revenue growth or margin growth, things of that nature. And it's not easy. Transitions are never easy, but ours been is a unique innovation-led transition, and we continue to feel very good about the fact, why? Because we are truly the innovator in our space, very well positioned in the market. I'll pause here. We'll peel the onion more later.

Yes. Maybe we can talk about this final leg, the sort of -- you talked about the third leg in the transformation. What's going to happen in this third leg? Why are we at this transformation? You gave a bit of history of how we got here, but it would be good to just double-click on that.

Well, look, I think the last mile of the transition is all about the [ true ] cloud transition. And I think that's a messy transition. I can be very candid about that one. And messy because the thing that -- to the technologists in me, the whole accounting part gets my head exploded. So -- and we just have to live with that, unfortunately. And the number of times I ask Mike those accounting questions, like, well, that's the way it is whether we like it or not. So no ARR is equal. So ultimately, we are moving to cloud. But the one thing to note in that transition is that while we are moving to cloud, we're giving up what is very highly upfront recognized revenue for a ratable revenue. So that's from an accounting point of view, kind of creates pressure on the revenue, overall ARR side of things, but we haven't skimmed on the bottom line. But at the end of the day, I think, look, there is -- that's the right thing to do for the company. But a very important to note -- part to note in that transition is that in our cloud business, which we've guided to be $1 billion of ARR this year, People think of us that sometimes the perception could be that, oh, it's moving left pocket to right pocket, the old customer moving to the new cloud. That's not true. Only -- and when we look at the TTM last 12 months, the cloud net new ARR that we added, 32% of that came from what we call modernizations, our on-prem customers wanting to move to our cloud, and that comes at a healthy flip ratio. 68% is non-modernization. In that 68%, 40% of that 68% is net new logos, logos like JetBlue, logos like NVIDIA. And the other 60% of that 68% is existing customers buying for new workloads. It could be Morgan Stanley bought us, bought PowerCenter from us 25 years ago, and they're an existing customer. But they want to go to the cloud in the context of, let's say, data governance, it's an existing customer going for a new workload. That's how we think about those. So a very healthy part of our business comes from new workloads or new logos. And that is not changing as we think -- look forward. The proportionality of that is not changing in 2025 because we feel in a growing market, we are gaining share and we are growing very, very well as much as we are also helping modernize our on-prem customers to cloud.

That's super clear. So there's an incredible traction and at scale, $800 million, just say, on the cloud alone is huge. So whatever percent of that growth is net new is incredible, probably dwarfs what many of these start-ups are doing. And you've got this incredible tailwind happening there. At the same time, Mike, maybe to you, the quarter had a lot of noise around it. And I think what would be really helpful in this session is to go through a little bit what happened, but really explain what's idiosyncratic for that quarter, what's enduring and maybe there is some things and like frame that for us in short term, medium term because you've got this incredible long-term tailwind that seems to be happening as Amit is talking about, but I think it -- there was a bit of murkiness in the last quarter that confused investors.

Yes. Thanks for the question, Sean. And I'll start with a couple of things. First, we have a complicated financial model. It just is a reality. And so a lot of things can move and a lot of things did move in Q4. We were surprised by a lot of things, surprised negatively by a lot of things, caused us to miss our guidance on most of our metrics, except for free cash flow, which we exceeded. And we guided to '25 below consensus, which is lower than we thought we were going to guide when we were thinking about what '25 would look like in October. I don't want to try and recreate the entire earnings call to try and reexplain everything. But let me go through some of them piece by piece with the punchline that I hope you'll agree is that it doesn't add up to a change in our strategy, to a change in our thesis, to a change in what we think is the present value of Informatica. It does shift out into the right some of the goals that we had set out for ourselves at the end of 2023. But the fundamental path we're on remains the same and the fundamental value, we think, is still very strong. So starting with revenue, which is impacted heavily by on-prem software accounting because we still have, as of today, $900 million of on-prem software that we are renewing, not selling, it's all end of sale, but which we are renewing. We saw the renewal rate come down somewhat in our on-prem business, about 2%. We saw the duration of those renewals come down by a few months, both of which add up to less upfront ASC 606 recognized revenue, which have a surprisingly large adverse impact on GAAP revenue when it flows through. We also had somewhat less professional services sales than we had expected. Professional services is a low, no margin business, and it's coming down because more of our services partners are anxious to do that business instead of us, which is great. We had a couple of million dollars of FX impact and it all added up to an unpleasant revenue outcome, which then flows to the bottom line. Now into -- and I'll get to ARR in a minute. In 2025, we think that those lower renewal rates, we forecast we pushed those through to 2025. We haven't assumed what we hope is going to happen, but we've taken steps to achieve is an improvement back to where we thought they were going to be, but we've pushed those through into 2025. We've also pushed the lower term length for self-managed or on-prem renewals into 2025. With respect to ARR, that's where the cloud economics are more -- are better exposed because that's a ratable business. The 2 big things that happened in the cloud and which we think we -- and again, which we are pushing into 2025 with our projections are, our renewal rate was a little lower, about 2%, not game changing, not thesis changing, but noticeable. And we've talked about this at some length, but our belief is that it was idiosyncratic. It was mostly execution related, and there are specific things that we can do and have done that we think is going to allow us to get back to a level we're happier with in 2025. But we're projecting low 90s, which is what we saw in 2024 in 2025 for renewal rates. The second bit is the mix of the gross new software we sell between modernization deals, the migration from our on-prem base to cloud and net new software sales, the 32% modernization versus 68% net new that Amit just mentioned. That was a higher mix of modernization than we forecast by quite a bit. And that's a good thing because that modernization has a great lifetime value to us. But the way the accounting works with the credits we give against the existing software, the double counting that happens for a period of time while the migration is taking place, it has an adverse impact on the total net cloud ARR that gets exposed during the initial period of those new cloud deals. So just that mix alone brought our cloud ARR growth down versus what we expected. And we have pushed a higher modernization mix into our assumptions for 2025 as well, which took down the net cloud growth number that we guided to in 2025. So that's an attempt to get to the structural things, Sean. There is a lot in there. But again, I would reemphasize that from our point of view, it is a shift out into the right of some of the marks we had expected to hit by 2026, but nothing in that list is game changing from our perspective.

Yes. In many respects, some of those are long-term good for the business. So think about the modernization, those things are actually great because they're going to be long-term enduring customers. You talk about this uplift multiple that happens when people transfer. I think most people will agree, it makes sense to give them credit. So how should we think about how those things will continue, why people are now choosing to modernize and how to think about the credit the multiple as well?

Yes. So when we modernize a customer from on-prem to cloud, they pay us more than they were. Cloud delivers more value to the customer than on-prem. We run the stuff, we patch it. We bear the hardware costs. the lift from on-prem to the cloud has historically been 2:1. It was lower by design in Q4, and we expect it to stay lower in 2025. We've guided to a range of 1.5 to 1.7 in terms of the uplift from what the customers pay us today on-prem to what they will pay us in the cloud. We are lowering that target or that expectation not because we're cutting price. We're allowing our sales force to address a broader amount of the on-prem cohort. And that is that part of our on-prem base that is using the software in a way or the software is priced in a way historically, that replicating that use case in the cloud takes less IPUs. And therefore, they just don't -- unless we want to charge a premium for the IPUs, they don't need as many IPUs as that earlier set of customers that we had modernized. So on the one hand, it looks that flip multiple, that uplift multiple is coming down, which if we could get 2:1 forever, we would, of course, but we can't. But the present value of those customers is very good. And we're comfortable now allowing that uplift multiple to degrade because we now know that when we sell that original modernization deal that it typically drags a significant amount of expansion sale upfront to do other stuff in addition to what they were doing with us on-prem. The utilization of the IPUs during the course of that engagement are very high and therefore, leads to expansion during the engagement. and it renews at a rate that is higher than the average renewal rate for our other cloud deals. So it all adds up to a very positive present value and allows us to be comfortable to lower that uplift multiple threshold to bring more and more of that base into the on-prem -- into the cloud platform.

Awesome. Now I'm going to shift gears and go back to some of the fundamentals here. What really drives the software company's valuation is really winning new customers and expanding existing customers. And you mentioned we have so many of them to talk about. So maybe could we share a little bit about what have been the nature of some of these new expansions or some of these new wins? Why do they come choose Informatica, JetBlue, maybe others? And why do you keep winning in this regard?

Yes. So I think if you piece part it into 3 categories. One is the one big reason why we've been able to win many more use cases are because by definition, we have many more products and that solve many more use cases. That -- if people don't know the Informatica story, people get confused with ETL back in the days, ETL is a very small subset of our business anymore. When you think of ELT, when you think of app integration, Master Data Management, which is 360-degree view of a customer of your supplier, of your supply chain, of your product governance and privacy and cataloging, many, many more use cases. So that has been a fundamental driver of a bigger TAM and more use cases. Second is, and I'll give some examples of that one is that customers inherently now have a very, very fragmented -- this market is so fragmented. And they -- to solve any use case, let's say you want to get a single view of a customer. You need an MDM, you need ingestion, you need quality, you need catalog, you need 100 things. The stitching together is so much high risk and takes more time and it just is high -- low ROI. And that's solving it on a platform like ours makes it easy. And the third one is that our relationships with both the hyperscalers running it on their platform and the GSIs is tremendously important to large customers as they drive transformation. To give you some examples, take JLR. JLR basically is Jaguar Land Rover. When the first deal they did with us is a massive transformation, the whole EV transformation they're doing EV battery. It was AWS, Snowflake, Informatica, TCS. They were going through a massive supply chain transformation to become an EV company, and it required a ton of analytics to understand the battery and all that stuff. And as you can imagine, it wasn't a simple deal. So they worked with us, and they landed on the first analytics use case with us. And from there, it grew into over the last 2, 3 years, that thing has grown into multimillion dollar more into now they want to do governance around it, whole supply chain optimization around it, and it grew from a single use case to multi-use case and multiproduct. They come like NVIDIA. Each GPU, as you all know, is a very precious GPU in today's world, right? Managing the supply chain for not only supply chain optimization, governance, who gets what, where are the chips coming from, where are they -- is a very, very massive project in itself, and we help them with that. JetBlue is a whole customer sentiment analysis to basically maintain their existing customers, make sure that they can get a higher value from them, but also opening up a whole new -- attracting more customers at different price models, things of that nature they experiment, including things like what entertainment they can give to customers and all that stuff, where will they swipe the card. So that became a customer sentiment analysis. So the breadth of use cases is significant. And of course, and like I said, all of these are multiproduct transactions and customers want to have relationships with like if you're doing this transformation for me, I'm an Azure platform. I want you to have deep technical relationship with them, which we do. And by the way, these GSIs working for me, they have Informatica practices. They have trained resources. Can you bring your resources and help them reduce the risk because we have technical architects and all that stuff. Those are the kind of use cases, complex, and I'm going to give you some examples on why we win disproportionately. And lastly, we future-proof our customers. People think of today's technology. Back when Hadoop was here, some of us were here and Hadoop was like the biggest thing since slice bread, and we don't even talk about Hadoop anymore. We helped customers carry their business logic from old into Hadoop world. A lot of our customers are running Hadoop workloads today can leverage Informatica Cloud to still run that same business logic into the cloud world. Same is happening for us in the world of AI. Same IDMC can run non-GenAI workloads, but we support RAG architecture, LLMs, vector databases. Customers are starting their pilots already for Gen AI projects on IDMC. So future proofing becomes very important for these enterprises. So those are a long answer, but how why we remain relevant.

I appreciate going into the detail. You've already hinted the AI, so we have to go there next. What are AI workloads you're seeing people do? Like what is someone actually building using AI and using Informatica?

Yes. So we -- I shared that in the earnings call, Sean, like about we track 100 customers because they're using, we can see the workflows across -- by the way, a variety of segments. It's not like it's technology companies, it's manufacturing companies, health care companies, financial services companies. And you see different kind of use cases. Customer sentiment analysis is an easy one right now, understanding the customer better, so we can service their calls. If I'm -- even in a financial services or a credit card company, things of that nature, I can manage that. Business process optimization is a big one. Customers are looking at their business workflows in a very different way, and they're looking at how do I optimize these workflows so I can increase the productivity of my operational workflow. That's another big one. Looking at things like net new revenue streams opening, whether it's basically, how do I monetize clinical trials to get to drug discovery better or how do I basically look at different transaction customers are doing to offer them new services or products. We're seeing some of that, that can be. So those are very, very different use cases we are seeing that are coming up on our platform. And the one uniqueness for us is that people think of GenAI and the only way GenAI is being used is everybody has got a GPT out there. That's not true. So it's true but not true. You need fundamentally a data management architecture to do Gen AI. We call it Informatica for GenAI. You need to ingest holistic data. You need to make sure data is of high quality before it goes into an LLM. You need to make sure that there is some governance on top. So you need all of those capabilities for Gen AI projects to get on board. And like I said, we support all of the LLMs and all that's what we call Informatica for GenAI. But you also need the Gen AI capabilities to drive productivity. So we have our own copilots. We have our own CLAIRE GPT. And I kind of shared that as we come to Informatica world, we've already announced the strategy for Agentic CLAIRE and we'll be basically unveiling a lot of the agentic strategy, and we'll demo a bunch of products over there. So that's how we also drive productivity of our customers.

Okay. And what about this Gen AI blueprint? So you put this out, it's going to be available in the cloud. Some people won't really know what they are. Can you explain what it is and maybe what it affects for go-to-market with these.

Sure. I mean we are the only one, by the way, in our market to have a GenAI blueprint with all of the critical hyperscalers and data platforms with AWS, GCP and Azure and Snowflake, Databricks and OTI and all of them. And the blueprint is that everybody has a different architecture for Gen AI. If you look at -- if you hear what Ali will talk about, he has his unique architecture of what he's talking about. If you go to Snowflake, what customers expect, what are we? Well one of the biggest reasons why we succeed is we are the Switzerland of data. And we're going to be the Switzerland of AI. If I'm a customer implementing Snowflake's AI for this use case, on Azure or I'm implementing database AI on, I'm purposely using different hyperscaler platforms for them. They are all unique implementations. And there's no one size fits all. Our goal is to support all of their AI technical architecture, so it's easy for our customers to go live with that. So our blueprint show customers what's the right reference architecture, what's the easiest way you can think about. And it also allows us to go technologically develop the right product between the 2 companies and allows GSIs to also go and put that blueprint to customers. So that's the purpose, and that's what we are seeing gain traction with customers.

What's the feedback been so far?

Gives them -- first of all, customers are dying for -- this is an area where in enterprises right now, so much noise. First of all, give them comfort that I can count on you. If I go hear database story and I come -- I have a platform from you by both of you can work together and I'm future-proof. And I -- first of all, I can begin somewhere. Remember in Gen AI. We all consumers have been 100 miles ahead than the enterprise on Gen AI. There is a serious lack of shortage of people who can do Gen AI projects. It's not like tomorrow, you can start a Gen AI project and boom off you go. So people need technology blueprints to be comfortable. They are learning, GSIs are learning so that even if -- by the way, today, most of those stuff is pilots, I can at least first kind of do the right pilot before I go to production. So it gives them a place to begin, comfort to put an architecture in place and also allows them to kind of scale without having to literally troubleshoot or code their way through every little piece of architecture.

And maybe just linked to some of the announcement, you mentioned how much of CLAIRE, for example, is being used internationally. I think it was 10%. How do you think about the international opportunity here? One, for the AI products you're building and two, just in general, international for Informatica?

I mean both synonymous. Look, I mean, any new technology first gets more used in the U.S. before it's going to international, no surprise, and it goes to the English-speaking countries first, and it's all tied to risk and how companies work. So I think any new technology follows us -- I have not seen it change much from a first U.S. and then international. I think on the mix of the business, do you want to add to that, Mike, on our mix of business international?

So it is 65-35 North America, the Americas and rest of world. And that's going to stay pretty stable. We see healthy growth around the world without a lot of geographic divergence -- it will be a little bit faster international than the U.S., but not a big mix. I mean what you're observing, though, which is important is that we're making that key feature, CLAIRE GPT, just available through more points of delivery around the globe, which is going to continue to drive win rates and the virtuous cycle of more people using it, therefore, it gets smarter and can do more for everybody.

Perfect. And Mike, maybe staying with you a little bit. One of the big things you guys did many years ago was the IPU. And it's now been many, many years. It allows you to interoperably use different modules on the platform. It's one different pricing unit. What do you think the impact has been this year? How do you think it's progressing? We talked about it last year, but what's the progress? Is it a home run success? Is there still stuff to do? Do you think the whole market has now moved to the IPU?

It's going very well, and it has been behaving the way we hoped it would. I used the term earlier, maybe I should have explained it when I did. The Informatica processing unit is a fungible token that you buy in advance on a committed basis. So you buy a certain number of IPUs for a multiyear period, use it or lose it. So it's not a direct drive consumption model where your bill changes every month or every quarter. You buy 100 IPUs, you have 100 IPUs to use and you pay us for that a year in advance every year. But the IPU is, therefore, usable across all of our services on the platform. And there's close to 40 right now that you can single pane of glass, literally point and click and use your IPUs for data ingestion or quality or catalog or push down or serverless. And what we see -- and the reason why the IPU is so valuable and so important is because it allows our customers to get the most out of the platform and to do so on their own. So if they are using ingestion and they want to use quality, they don't have to come back to us and negotiate a new SKU and a price, how much am I going to pay for quality. They just use the IPUs they already have. And once they've gotten to close to full utilization of those, they just call us up and they buy more IPUs because the price is very clear. And what we see in the data is that when customers are with us after the first 6 months of their initial land on the IDMC platform, on average, they use 3.6 services out of the 40. They need 3 to 4 services to do their initial use case. But after they're with us for 18 months, they use over 7.5. So they double the amount of services that they use in that first 18 months, which -- and that's in virtually all cases without a new sales cycle, without some sort of high-touch engagement, it just makes it so easy to expand yourself across the platform and get value from it. And by the way, we're seeing that the utilization of those IPUs in terms of consumption is healthy. We're seeing the interim expansion of that is healthy. So it's worked out really well.

Maybe on the topic of -- you guys are one of the best sort of examples of a multiproduct platform. There are a lot of things you can do in Informatica. You gave some stats just there of some of the progression for the IPU-based business. What does it look like overall? And what does it look like for a customer to start adopting 10-plus products or something? Do you have any like that?

Yes. So the only pricing models we offer right now are IPUs and master data management, MDM records. That's all we sell. So Master Data Management, the Customer 360 is, by nature, a per record business. Everyone in the market sells it per record. So I don't want to master 1 million customers and we sell you 1 million records or what have you. And that's the only reason why it isn't on the IPU model because that's just the way the market wants to buy it. So everything else is either IPUs or MDM and all the legacy stuff is fading away as we renew or migrate those contracts to the new model.

And Sean, not every service is equal also. So I talk about 40 services, which are not including MDM. But like, for example, we have CDAM, data access management, right? The utilization of that service is not the same as, let's say, ingestion. So I think there's always like a same -- so I think -- and CLAIRE GPT is a service which is in its early days. And so when you need underlying IDMC. So of the 7, what we want is definitely more services, but we want more compute-intensive service because that creates a ton more stickiness also.

Maybe then related to that question, how big concerning customers get? What do you think the pro [indiscernible]?

It actually -- well, great question. I don't have the top of my mind, but take a customer like Takeda. Takeda actually won the innovation award from us at Informatica World and Barb, who was the CTO was on stage, and she talked about how she began with a small use case -- and literally, she's got the whole platform, including our data marketplace, our CDAM, they have master data management. they have governance, they have catalog. They obviously have all of the integration capabilities. Now again, in integration also, they have 3, 4 flavors. It's ETL, ELT, mass ingestion, serverless, I can go on and on and on. Depending upon the customer's use case, if I'm on, let's say, Databricks, you may use serverless. If I'm just using, let's say, a warehouse use case of Snowflake, I may be fine with ELT. So a great example of a customer that has used many services, still not all 40 services, but probably like 15 and those 15 are driving multimillion dollar ARR for us as an example. So it's not necessarily just a number of services, but we want -- we know that we can go from A to B to C in terms of driving multiproduct once they land on a place because cost is less for them to go from A to B.

And if I may interject, that's one of the reasons, Sean, why our net retention rate is so attractive. It was 124% in Q4 and it's going to vary a lot. So I don't want to guarantee a particular number, but anything 120% plus, I think, is really good. And that's because the platform has so much you can do with it that you can land and expand. The pricing model makes it very easy to do so. And the nature of our customers, generally large enterprise customers have a lot of footprint inside every organization that we can move to once we get that initial foothold.

What do you think the main driver of expansion is going to be? Is it going to be just data increasing, driving more compute? Is it different modules being used? Or is it just going to like different sub teams within a massive company? I know they're all slightly different, but what do you think is the big one?

I think the good news for us is that we have so many vectors in which the usage can grow. Think about it this way. I think it's not just data growth. The data growth, I've never put a data growth as a fundamental driver because, look, I always look at click stream data, the amount of data that goes mostly junk, right? If you re-loop it, basically, it's like 100 to what really matters within a company. So it's about -- for us, what matters more is there is every use case within an enterprise, which is digital is a data-driven use case. That is fundamentally growing. The complexity of use cases is a big thing for us, right? Because the more complex it is, customers need help from a platform and a high integrity player like us. The more -- by the way, future proofness is very important to us. As customers are going to Gen AI, they want future proofness of I can run this and I can run that or that one. That is a very important one. Fourth one becomes important, can this company with all these use cases run it at the scale I want. If I'm a Morgan Stanley, it's not just a functional I need stuff, but can it pass my security teams? Can it run internationally? Can it run that -- and will it -- do they have the scale to support me? Those things become very important. The customers very quickly adopt some use cases because those require. All of those things become very unique thing versus just functional capability. So that's why I feel very good about the fact that we have multiple routes to drive expanded usage.

And what do you think about the world today in the data space? Is it going to get more fragmented or less fragmented in the next few years?

I think it's supremely fragmented already.

I kind of sometimes like it's a crazy long tail. And I think customers are now finally like really struggling with it. I mean, look, my core belief hasn't changed. I think -- and if you look at any market, at some point, fragmentation is counterintuitive. I think we can all argue it's always due to cheap availability of capital. If you take the number of VCs, everybody wants to invest something in one space before you know there are 500 companies running around. And I think at the end of the day, I'm hearing it loud and clear from customers. During the [ hay day ], everybody could experiment for 100 things, they want to consolidate. By the way, you're seeing the same thing in the world of AI, I look at the fragmentation over there. But consolidation naturally happens because not all of them can live to help our customer scale, and that's what people want derisking.

Fred [indiscernible] ...

I want to remind folks of what you may or may not have seen in our tech talk as we called it in December, where we went into a deeper dive in some of our products and our value proposition. And one of the things we emphasized was that our core customer base is not the 300-person Silicon Valley company that is standing up their first data warehouse and can do it in a relatively simple manner. They've got a few different sources and they've got one target. That's not our customer base. Our customer base is the large, generally global company that has a very complex infrastructure and lots of complex challenges already. And those customers are multi-cloud. They're also hybrid, but they're multi-cloud. And every survey, every piece of data you look at, those customers say they're getting more multi-cloud, not less. That is directly to our strength. We help you knit together the data that goes into your multiple clouds. We help you pick it up out of this cloud and move it into that cloud if this cloud gets more expensive or this cloud doesn't have the features that you want that you have over there. We future-proof you against changes in formats and different styles of data movement because that's what we do every day. Those trends in terms of fragmentation are getting worse as it were to our benefit.

By the Uber wasn't a customer and they were a small company, Uber is a customer now for us.

That makes a ton of sense. Take some questions quickly. You can just read out. [indiscernible].

Yes. So we absolutely intend to modernize as much of that base as we can. There's still $900 million of it, and we've got a long ways to go. In 2025, we've based our forecast and our guidance on the assumption of about 30% of our net new ARR is going to come from modernization, moving the on-prem and about 70% is going to come from net new workloads. So you can do some simple math based upon the NARR that we're projecting, and we're still going to have a lot of on-prem yet to migrate at the end of '25. So it's going to be a long tail, and we're going to moderate it in a way that we think maximizes the long-term value. We're not going to do anything artificial or particularly punitive to try and force customers with a big stick, at least not -- that's not on the horizon. On the cost side of things, we took actions in 2023 after we announced the end of sale of our on-prem products that rationalized the cost we had against those on both the R&D and go-to-market side. So we've taken that cost base down to a pretty attractive level, and we're investing now in the cloud, both go-to-market and R&D. So the low-hanging fruit of the cost reduction of the cloud-only strategy, that was harvested in 2023 and 2024.

Maybe one thing before we wrap up. What are you most excited about for 2025?

Can I use one thing just to [indiscernible] through so much and in a mixed model, it can be very complicated. I want to simplify. We have $900 million of on-prem ARR, which we are basically not selling into. It is churning at a natural rate and whatever is churning at more than naturally will move into cloud, which we just talked about modernization. But what I'm super excited about the fact is that while we have that is that we have a $1 billion what we've guided to this year, cloud ARR business growing at 25% with an NRR of north of 120%. We ended at 124% with all the intrinsic tailwinds that we talked about. So I think in the from to mix model ARR, rev rec, everything gets lost, I look at the business 2 ways. This is a business that has all the tailwinds of multi-cloud, AI, scale, platform, Gen AI that we feel very good about, while we migrate that and that we acquire a lot of new customers. I feel very good about that. But to the question you're asking, I feel this is a year where, yes, we have some work to do, no doubt about that. The renewal rate on the cloud that we've got to recover back. But I feel very good about this and the innovation that's driving that. We still remain the #1 in our markets -- we still remain the only company that has a platform. We still remain the only one that has blueprint with all the hyperscalers, the cloud platform. We still remain the only company that can support in our market customers of the large enterprise scale. And we feel very good about that, while we still do the from tools that can sometimes my head explodes on revenue accounting and create complication. But there is no one of that size, scale and moat that we have around us.

Anything you'd like to add?

I couldn't add anything to that.

Amazing. Well, thank you both for coming, and thanks again for coming here.

Thanks, Sean. Appreciate it.