Hewlett Packard Enterprise Company (HPE) Earnings Call Transcript & Summary

Good day, and welcome to the Executive Deep Dive HPE GreenLake Cloud webcast. [Operator Instructions] Please note that today's event is being recorded. I would now like to turn today's webcast over to Meta Marshall, Managing Director at Morgan Stanley. Please proceed.

Great. Thanks, everybody, for being here today. We're delighted to have Fidelma Russo to answer questions for us and give us kind of an overview of the hybrid cloud business. Fidelma is the EVP and General Manager of HPE's hybrid cloud unit and is the company's Chief Technology Officer. She joined HPE in 2021, as CTO. Prior to that, she held notable roles at VMware, Iron Mountain and EMC in addition to kind of quite a bit of others. So will be able to kind of answer all of our questions today. So with that, Fidelma, I'm going to hand it over to you to kind of walk through some slides and overview of hybrid cloud [ development ].

Thanks, Meta, and thank you for having me here today, and hello, and welcome to all of you who are joining today's webcast. I know your time is valuable, and I appreciate you being here to learn more about HPE GreenLake cloud and why it's important to HPE's long-term strategy. Now we've been on quite a journey. We've made a lot of progress in a relatively short amount of time. And so why don't we just get started with why GreenLake is a cloud and the best cloud for a hybrid environment. So when I think about GreenLake. I take the analogy of the public cloud and the rise of the public cloud in the last decade really revolved around a certain sort of aspects that have made it easier for IT organizations to do their jobs. Thanks, Meta, and thank you for having me here today, and hello, and welcome to all of you who are joining today's webcast. I know your time is valuable, and I appreciate you being here to learn more about HPE GreenLake Cloud and why it's important to HPE's long-term strategy. Now we've been on quite a journey. We've made a lot of progress in a relatively short amount of time. And so why don't we just get started with why GreenLake is a cloud and the best cloud for a hybrid environment. So when I think about GreenLake. I take the analogy of the public cloud and the rise of the public cloud in the last decade really revolved around a certain sort of aspects that have made it easier for IT organizations to do their jobs. The public cloud is still filled with data centers full of computing, networking and storage infrastructure in various locations, but they provide services in a consumption model and sometimes in a shared tenant model and sometimes in the reserve model. And you access these services with a platform where the catalog is displayed and customers subscribe to the services and they pay for all of them in OpEx. Now HPE recognized early on back in 2018 that not all workloads would go to the public cloud. And there are many reasons for this, which we all know, latency, data gravity, compliance, security. And we recognize that there will continue to be workloads in data centers or at the edge or in colocation facilities. In fact, today, I think none of us would argue that the word is really hybrid and that workloads need to go where their data is or where the right place is for those workloads, whether that's in the public cloud or in other private cloud locations. We also recognize that customers really valued the cloud experience that they got from the public cloud. And the customers really didn't want this bifurcated experience between what they got in the public cloud in terms of the platform experience, in terms of the consumption experience, in terms of the fact that the public cloud vendors manage the services for you. And so that was such a radical departure from what IT organizations experienced within their own data centers. So this is why we embarked on the journey of building GreenLake cloud. Now we did this with a combination of organic and inorganic investments over the last 7 years. We, first of all, recognize the world will be hybrid. Then we recognize that customers really value having an OpEx option to pay, a pay-as-you-go option in addition to their traditional CapEx options because on GreenLake cloud, you can actually choose to pay either as you use this option or in CapEx and this is really ideal for some industries who have large CapEx budgets for instance, real estate investment trusts, where the IT budget is a much smaller portion of their CapEx spend. But where from a tax perspective, OpEx spend is not ideal. And so on GreenLake cloud, while we introduced an OpEx a way to pay, we also allow you to pay for your infrastructure on-prem in CapEx. Then what we said was we needed a common platform for GreenLake cloud just like in the public cloud, which would allow customers to browse our catalog, consumer services and the real key piece of this is so that IT operators only needed to go to one place in order to manage all of their infrastructure. The next step that we did after that was we put all of our infrastructure and the management of that infrastructure on that platform in that catalog. So you can think about it as networking, Aruba Central, which is the management application that manages all of our Aruba assets is on the GreenLake platform. We follow that by our Alletra MP storage management, putting that on the GreenLake platform and our compute management, compute ops manager onto the platform. So now you can go to a single place and whether you are doing storage, networking or compute just like in the public cloud, you access them in a single platform, and you can cross-reference between them. In addition to that, what became critically important over the last couple of years is the emergence of private cloud instances. And here, in GreenLake cloud, we have 2 options: one, cloud connected and available on the platform along with your other services and the other air-gapped, for reasons of sovereignty, security and that is becoming more and more important as we navigate the current world. And then not only is it sufficient to have your infrastructure and your private clouds, on your platform. But if you look at what the public cloud vendors did and how it's resonating with customers, they really also put the manageability of all the assets on the platform. They put things like financial operations on the platform. They put things like data protection and cyber resiliency on the platform. And so we embarked on a series of software acquisitions. Morpheus Data, which we just completed last year, which brought us a cloud management platform on GreenLake allowing us now to manage your assets, whether they -- in a common way, whether they're on-prem, whether they're in the public cloud, whether they're HPE assets or whether they're multi-vendor, because we recognize the world is multi-vendor and multi-cloud. The second one we did was OpsRamp and here, it's very important that when you have all of these assets running on GreenLake cloud or other clouds that you have a single way to observe the behavior of the assets, identify issues and remediate either autonomously or with assisted agents to remediate problems. And so OpsRamp is also on GreenLake cloud but also manages multivendor, multicloud assets. And then finally, last but not least is Zerto, our data protection and cyber resiliency offering. And here again, multivendor, multicloud and providing protection across the portfolio. So what this allows customers to do, these software assets are very sticky. And what it allows our customers and our service providers to do is to either have the estate managed by HP, have it managed by a service provider or have it managed by an internal service provider within a customer environment, just like the public cloud. So to summarize, we embarked on a journey. We embarked on adding the things, first of all, the consumption experience, second of all, the platform; Third, the cloud operating model, the choice of how you operate and manage your systems. We did this organically and inorganically. And now we have a complete offering of a hybrid cloud with GreenLake cloud. And this is really resonating with customers. Today, as I talk to -- I talk to a lot of partners, I talk to a lot of CIOs. And what we see in the market trends are high degree of interest in making sure they can manage sovereignty, making sure that they are looking at value in the run times and many CIOs are looking at modernizing their estates to manage their runtime costs. And then looking at how do they bring GenAI across their infrastructure and across all of their operations. And this is a perfect time for GreenLake cloud because we have multiple ways of helping a CIO manage these, in some ways, competing issues within their budgets, which are really not increasing. We have our HPEFS asset management recycling capability, which is a way for customers to recycle their prior generations of equipment and modernize their environment. We have run time choice with our Morpheus Data systems and our own virtualization and container technologies to allow them to get -- to have a more value-oriented run time environment for them. And we also have all of the cloud operating models to make it simpler for them. And then our private clouds, we enhanced our Private Cloud portfolio last year with the addition of Private Cloud AI in addition to our direct liquid cooled servers. And so we have a large portfolio on GreenLake Cloud to help you implement very quickly and bring value in how you're deploying GenAI within the enterprise. So a few factoids around how we've been successful with GreenLake Cloud and why we feel confident that we're well aligned for the future. Today, we have over 41,000 customers using GreenLake cloud. Last quarter in Q1, you saw that the ARR crossed $2 billion for the first time. We've had a CAGR on GreenLake cloud of 45% year-over-year. And a net dollar retention of over 110%, which is world-class and so which demonstrates how sticky this GreenLake cloud can be. We also see much higher attach from our storage devices to compute within GreenLake cloud. And so we're very confident that we are on the right path, and we're aligned for future trends. So thank you for allowing me to have some time today to just walk through the journey on GreenLake cloud, share some of our insights into why we feel that we are the right hybrid cloud for the next-generation IT. And I'm looking forward to diving into questions with Meta. So I'll hand it back over to her. Thank you for your time today.

Great. Perfect. Fidelma that was a great overview. I guess just maybe to kick off and just to start, as a reminder, for everybody, please enter questions kind of into the portal. I'm going to run through a bunch of questions, but happy to ask anything additional people want. I guess just given that this has kind of been a segment for around a year. Just kind of how is your perspective on that business changed just having it be part of hybrid cloud segment?

Yes. I mean, we created the segment because we had -- we finally had, I think, enough momentum in the different pieces. And the piece that I was excited about when we formed the segment is we now have the opportunity to really build integrated solutions for customers and make sure that we're bringing these pieces to bear. We're at the right time as customers have these conflicting issues within their budgets, it's add more stuff, take more money out. And I think with the GreenLake portfolio with Alletra MP with our private clouds, I'm more bullish than I was when I took the job at the beginning of the year.

Got it. As you mentioned, kind of as you were closing your presentation, hybrid cloud has had quite a meaningful kind of bounce back and continued growth rate to continue through fiscal year '25. Seemingly led by Alletra but also with GreenLake. Can you just kind of give a sense of what the competitive environment is for GreenLake because it sounds like a very compelling value proposition. Just want to get a sense of that landscape.

Yes, it's a great question. And so having been in the IT leader shoes myself, I got a great appreciation of many of the things that IT organizations consider when they're looking at kind of 1 year, 3 years and 5 years. And for GreenLake, especially, it's really the time is now. GenAI has actually really had people take a step back and look at where they want to put their workloads and especially around their data. Data has become more and more important. And customers are seeing data as number one, a competitive advantage. So they don't necessarily want to share it on the public cloud to feed other models. Number two, it's very hard and expensive to move data. So that's the second piece. And so what they're really looking at is what do I want to do in the right place? What do I want to do in the public cloud? What do I want to do on the private cloud or on-prem. And that's really driving a lot of conversations on GreenLake. And once we have that conversation, we then move to, okay, if I'm going to put a combination of generative AI workloads, maybe some in the public cloud and some in the private cloud, how is my team going to manage it in a consistent way. And that's where our operations model like Morpheus and OpsRamp come into play. And then the third, as we're having these conversations, there are some organizations who are really able to kind of find additional budgets for GenAI. And so the conversation doesn't change. But there are a lot of organizations who are trying to manage those budgets. And so they're trying to reduce ballooning run time costs while investing in GenAI. And this is where we really see the private cloud aspect of GreenLake with our -- with the fact that we can use multiple run times, our own run time with VME. We can coexist with VMware, a great partner of ours and allow you to look at which workload you put on, which run time. And so from a -- so from a hybrid cloud perspective, I think, we are ideally suited because of our strength in having a platform that's hybrid and our strength in the on-prem infrastructure. And so that's kind of -- that's really when we are in a competitive situation, that's -- we have the consumption model. We have a platform that looks like a public cloud, and we have a wide variety of services and you can either manage yourselves or not, and then we have the financial services that will help you refactor your environment as needed.

Got it. That's helpful. I mean, when it comes to kind of that private cloud business, what are -- where are you seeing the most traction there? Are there key verticals that are kind of showing the strongest adoption of that and kind of the GreenLake portfolio?

So we have a very broad private cloud portfolio, and we're -- and so if we can walk through it because of the different segments, so we have private cloud enterprise, [indiscernible] say is our large data center. We manage it for you. Then we have private cloud business addition. So more for data centers, but maybe smaller footprints and some of the edge. And then we have Private Cloud AI, which we codeveloped with NVIDIA and really targeted around the GenAI workloads. The first 2 are general purpose for all sorts of bare metal container and VM workloads. So where we're seeing things and 1 of the big adopters in when it's cloud connected, we see it across the board. We're seeing it interestingly today, we're seeing a lot of companies that have edge locations like oil and gas companies, like retail companies, as they're relooking at how they're going to redo their IT. We're seeing a lot of interest in a centralized private cloud enterprise and then a ton of edge locations with private -- with PCBE. And if you can -- if you imagine, that's a lot of equipment and it's a lot of management. And the way that they're pulled towards it is, it's consistent management across all of it. So this -- we're starting to see that in -- across the board, insurance companies, finance companies and companies with lots of edge locations. If we take Private Cloud AI from a GenAI use case, there, we're seeing a number of different things. We're seeing IT organizations in big financials, but they're developing -- they're not developing the financial GenAI bots there developing chatbots for internally. They're being used for service helpdesks. We're seeing energy companies doing the same thing. So we see a swath of horizontal use cases there. And we also see a lot of interest in air-gapped private clouds and air-gapped private clouds, especially for defense sectors, for companies that are in Europe, especially in sovereign areas who really want to have the equipment and the management managed in country on their soil. And that's coming from government agencies, defense agencies, state and local government, and it's the same here in the United States.

Great. That's helpful. I mean, just in terms -- you just mentioned kind of some of that air-gapping. But are there other ways for you to target customers more or product enhancements that would kind of help take advantage of this opportunity or could kind of get people over the line?

Yes. I mean I think we have -- so from a product perspective, there's air-gapping. We're seeing the run time modernization and getting more value out of the budget that you have for run times, how do I get more value out of it. And that's where we have a ton of interest in VM Essentials. And we got VM Essentials as a part of the Morpheus acquisition. And Morpheus for everybody on the call, it can manage, it provides you a consistent management plan. for any of your run times, so Red Hat, VMware, Nutanix, our own VME and our own containers and ones in the cloud. So that is a great interest to customers because they can get a consistent model and then underneath, they can pick and choose the variety of run times that they have. And so that is the second one. And then the third one is are some of the enhancements that we're bringing to bear with Alletra MP and our X10000, which we introduced back in the end of last year, ground up design for a key value store object. And so what does that mean? It means it's fast. And it means you get rapid restore. We've also embedded tagging and vector enhancements for AI workloads, working with NVIDIA on RDMA. But really, what does that mean? It means that we now have in our portfolio, a storage device that is built for the next-generation AI workloads. And so as customers are looking at, they have different GPU farms, maybe they have -- they're running their workload in the cloud, but they want their data on-prem. So they have proximity from there and they put the data on-prem on the next 10-K. So we have a lot of innovations that I think we're bringing to bear that are driving these customer conversations.

Got it. I would assume that, that kind of represents most of the road map for kind of the next 12 to 24 months. But are there other areas like AI specific where we should be thinking of kind of more innovations kind of coming out over the next couple of years?

Yes. I mean, we -- as being the CTO, we spend -- I spend another part of my life, watching what's going on. I think we're at the start of the AI evolution and transformation. And we've seen a lot of progress with the CSPs, with the cloud vendors, but enterprises are really just starting. And the good news about that is I think we will see more and more innovations coming quicker. So the DeepSeek announcement back in the beginning of the year really demonstrated to all of us that this part of the industry is going to go the same way we've always gone, okay, which is people will figure out ways of making software more efficient, taking algorithms, making them better. And that's only good for all of us. And so we -- within our portfolio are looking at what we can do to optimize in HP Labs. We're looking at how do we see these things evolving. And then when we look at our software stack and how we're integrating it with NVIDIA, we are also looking at how do we make sure that we really decrease that time to value for our enterprise customers. So when I talk to enterprise customers, especially around GenAI, it comes out in 2 places. One is we don't have skills to really take all of the different pieces of open source, put them on servers, optimize that stack. And then another set of customers who are -- we may have these skills, but GenAI is so broad across our enterprise, and we want to make sure that those particularly skilled individuals are really working on the biggest dollar driver for the enterprise, but not necessarily on building productivity for everybody. And so what we see is we are doing a lot of innovation in the end-to-end stack to make sure we do life cycle management to make sure that we integrate new models quickly. To make sure that optimizations like that were done in DeepSeek, which we don't use, but the engineering behind it, making sure that, that's built into our pipelines. And so there's a lot of innovation, I think, coming in software over the next 18, 24 months that will really change the value for GenAI and help adoption over the next number of years.

Got it. January seems long time ago at this -- for all of us. It's been a long year already.

It's been a long year already.

Yes, exactly. I wanted to dive deep kind of on the VMware comments because you had noted kind of that being an area where you were seeing a lot of requests. Just what are you doing or how are you taking advantages of kind of customers being pretty discontent with VMware and kind of some of those price increases right now?

Yes. So I mean that's where sometimes you -- last year, when the Broadcom changes happened, and they've continued over the year. But when the large change happened at the beginning. We had a number of customer roundtables. And one of the pieces of feedback we got was -- while HPE has always been embracing choice for customers, we've embraced choice across the board, and we have a large business with VMware. They basically said to us continue to embrace choice. However, we want you to have an option so that if we continue to buy HPE, we want to have a run time coming from HPE. And so our initial plan was -- and when we were in -- we procured Morpheus, we already had plans on the table, but that just accelerated it. And our initial plan was just to offer this alternative run time, both VMs and containers, offer that within our private cloud portfolio because we were already working with Morpheus as a partner. The feedback that we got at that time was -- that's really nice. However, we have a whole infrastructure brownfield, other people servers, other HPE servers and we need you to also offer it to run there. So that we can transform our landscape. And we pivoted it, and we are highly committed to -- first of all, it's going on in our private cloud. But second of all, on HPE servers, on other heterogeneous servers, and supporting other back-end storage to help our customers really do this transformation. Either today, I think, I see it in a couple of ways. People who are doing it now, and that's what's driving our edge business, is because you've got a lot of sites with a lot of budget and really looking at how do we transform and make this easier and have more value. And then we see it now, and we see in a couple of years' time, the way still continuing. But what customers are doing now is they're testing. They're evaluating and they're really figuring out how they move certain applications over the next number of years.

Okay. That's helpful. And then, I mean, maybe just kind of turning to how Juniper would kind of fit into GreenLake. What synergies do you see between kind of the Juniper's portfolio and the GreenLake portfolio and just kind of how would it enhance kind of the overall offering?

Well, I mean, clearly, we're very excited about the Juniper acquisition and we plan to leverage Juniper on GreenLake. So it will be on the GreenLake platform, so you can access your devices, your Juniper devices in addition to your Aruba devices, your storage devices on the platform. We are excited about the data center switching side of it. They have some software-defined networking assets that will be very complementary to the portfolio. And so we are -- we think it will bring a lot of value to our customers and to us. And then the whole piece around networking for AI and kind of some of the switches that they have, it's really critical as we continue to drive AI infrastructure for our customers.

Okay. Maybe kind of continuing down that AI path. Just how does the NVIDIA partnership kind of contribute to what you're able to offer to customers. I think we delved into some of the GreenLake AI offerings, but just kind of outside of that.

Yes. So we have a very tight relationship with NVIDIA. Just to give some examples, we take their software code trees when they're just in test. And so we are jointly integrating their NIMS trees and their NVIDIA AI for enterprise code trees and into our AI essentials. And so we do this every day and test in both in NVIDIA Lab and HPE Lab. So the relationship is very tight in the field, which I think is really critical is we spend a lot of time doing roundtables, joint round tables, joint customer calls, making sure that we're really honing the use cases. And so that's in the enterprise side. And then on the kind of flexible solutions, which is with our liquid cooled systems and our storage. We're also working with them on some of the larger transformations within certain customers. So it's a very, very tight relationship, and it really got started back in June of last year.

Got it. I mean just maybe in terms of how you're seeing customers spend right now. Just how are you seeing them spend on updates versus spend for AI? And does that -- are there emerging kind of industry-specific use cases that you're kind of seeing direct more spend than others?

Yes. I mean I think there's very few of us at, again, extra budget every year. And so you're constantly trying to juggle this. What we're seeing and I had -- we had a customer advisory board maybe this time last year. And we did a survey there of customers and their AI use cases. And this is a group of customers who are enterprise customers, but some of them are like scientific research. So really one end of the spectrum to industrials, energy companies, to retailers, so a very, very broad swath. We did an exercise with them around how many GenAI proof of concepts are you running, okay? And the math and how many of the -- well, how many pilots experiments, how many get to proof of concept and then how many are you moving on to really implement. And the math that came back last year was 1 out of 5 moved through that pipeline. We just did the survey again, slightly different group of customers, but same broadness in terms of industry. And now we're seeing 3 out of 5 and so -- and the -- and what has happened, I think, is that first of all, people are getting way smarter about just narrowing the set of use cases they're going after first versus I'm going to experiment on everything. So there's more chatbots work, there's more services, there's more legal documents and contracts. I mean these are great use cases and they save a ton of time. And so -- and they're very horizontal. And so what we see is you can see we've gone from 1 out of 5 to 3 out of 5. So 20% to 60% in a year, which gives us a lot of confidence that people are really going to start to deploy and make it part of their infrastructure. And then what I think you'll see is once people get used to using it every day. And for instance, I run a large development team and have a lot of software development engineers. We use 2 GenAI tools to help us in software development and initially, the uptick was slow and -- but then once people kind of get used to how do we work with these new tools, how efficient are they? You actually start to see people saying, I'll never go back. And I think that you will see that across enterprises, once they understand how much time is saved and how -- not easier their job is, but like how they have time to do other things. I don't think anybody's job kind of gets easier. You just get to do other things.

Okay. Okay. That's helpful. Just in terms of -- maybe kind of going back to the GreenLake question of -- are you seeing kind of GreenLake or kind of cloud use more proof of concepts and then once they're kind of fully deploying, they're kind of coming on-premise? Or just kind of what is that GreenLake versus when that...

Yes, it's a great question. And so we see a couple of things. And I can give you our example internally first, it's probably a good one and then a couple of customer examples as well, to just kind of illustrate the point. So internally, we had -- like every company, a big push to adopt GenAI across the enterprise. And the first thing you want to do is to adopt chat. And so we started on a public cloud, and we started maybe -- we actually started 18 months ago. And then our uptick was slow. Is the same -- it was quite interesting. And then once people got used to us, the uptick started rising. When we came out with the on-prem with Private Cloud AI. Our IT team actually looked at it and said, we have 2 use cases. First of all, we cannot afford -- we can't afford all 60,000 of us to be running this on the public cloud. And so when they did the TCO of on-prem and the TCO in the public cloud, the benefits were huge. But that is just 1 factor. The second factor is that there is a lot of data that we don't want to go on-premise, and this will lead me to customers' conversations. And that's compliance data, merger and acquisition data, specific customer contracts that have to be really kind of -- can only be seen by certain people. And so there's a number of places where you need to have an on-prem instance in order to make sure that everybody can kind of do their jobs and benefit from this technology. So that's been our experience, and we've been bringing more and more of that on-prem. So now we go to customers. And what we see is some customers want to run on-prem because they believe that it's a better TCO and it has a cloud experience. We have those. We also have customers who look at the classification of their data, okay? And they will run POCs in the public cloud to kind of really hone in on the use case. And then based on the classification of the data, they will decide whether that use case stays on the public cloud or whether it goes on-prem. And so a lot of this comes back to where is my data, who can access my data because even all of those questions are really the most important pieces on this.

Okay. That is very interesting. And so just how should we think about, with all of that with eventually kind of some of the spend comes back on-premise. Just how should we think of the pull-through as part of the server refresh you guys have been experiencing?

Yes. So -- it's -- so on the server refresh side with, for instance, the -- we've got 2 sites going on. We've got the generic, x86 ProLiant Gen12 refresh going on. And actually, it is tied to somehow how people want to free up their budgets. For instance, if you have a Gen12 ProLiant, it can replace 7 Gen10s and also reduce your energy consumption by 65%. And so in the TCO, you have less footprint, easier to manage and basically a much better and much lower energy build, freeing up budget for you to invest in your GenAI infrastructure. And so that's 1 place we see it. The second place we see it is just aside from GenAI, that is also another place where as you're looking at your run time, getting more value out of your run time. Obviously, you have less cores and so therefore, you can spend less money on the software that's running on the Gen12. So I think there's a big tailwind coming behind the server refreshes that will benefit people's budgets to allow them to free up spend on GenAI. And then, of course, we announced a whole series of updated GPU servers. We're in the middle of the transition. And we have a lot of customer interest in those updated GPU servers.

Got it. And so about to kind of hop to the storage piece of the conversation, but maybe just as a bridge and as we close out kind of the AI conversation, just any other products or solutions you would expect to have to kind of address that AI opportunity within storage.

Yes. I think one of the acquisitions we did years ago gave us the Ezmeral Data Fabric and sometimes timing is everything in this industry. And it might have been a bit too early for its day. But now we're embedding that data fabric in all of our private cloud and also working with customers on using it to help them manage -- get -- manage their data consolidation for generative AI. So the thing about GI, which we all know is it feeds on data, okay? So 2 things. You have to have the data and the second thing you have you have to feed in GPUs as fast as they can take it, okay? And the problem with people's data, if anybody has been through doing kind of a master data management project, all of these, I'm going to take these big -- I'm going to build a huge big data lake. Many of these projects for companies have taken a lot of money, a lot of time and most have failed. And so with the Data Fabric, you don't need to move your data. This allows you to basically connect what I call all these puddles of data, whether they're on-prem, at the edge or in the public cloud. So again, it's multivendor. We don't care what you have, and we support many standard interfaces filed -- file object streams. And so basically, you can now install this and now you have a single way of looking at your data, which is 1 important step in making sure that you are ready to implement GenAI because many people get stuck even before you get to GenAI, the first problem is how do I get out my data. And so this is a piece in our portfolio on GreenLake Cloud that is actually having a resurgence because of GenAI.

Okay. So maybe diving into storage, the Alletra MP portfolio has been quite successful. Just how should we think of that relative to kind of the other Alletra products? And just how big of an opportunity do you see for the refresh of the storage business over the next couple of years?

Yes. It's -- so we embarked on the Alletra MP architecture a number of years ago. As everybody knows, we've had a lot of block architectures within our storage portfolio, and we are on a journey to really bring them all on to Alletra MP. And so -- and we've been quite successful in doing that as we've been building out the Alletra MP block side of the business. And so we see the refresh opportunity, especially within the installed base as critical to our success over the next number of years. I say it in store. I've been in storage for a long time. It's kind of like no tile left behind. And what customers get on the Alletra MP side is a standardized hardware platform with block. We're not going to add file a little bit of file to block. So you have additional features over what you had in the prior generations. And we think that this is -- we've already seen quarter-over-quarter increasing momentum on Alletra MP. And actually, in Q1, we -- in our flash arrays okay? We finally flipped from less than 50% of the flash arrays that we ship to more than 50% of the flash arrays that we ship with Alletra MP. So we're starting to see that in the installed base and with net new logos. So it's critical for the growth that we need in the storage business going forward.

Got it. We've talked a lot about block. You've been very successful over the past few years, just kind of with the third-party IP on file and object. Just how should that help kind of the margins of the business as you incorporate kind of the block refresh? And just what is the opportunity for upsell of other products as you kind of go through that refresh?

Yes. So I mean, if you look at the storage market, it comes down to block file and object and then data protection. I mean, we've lots of other things around it, but those are really the core pillars. And our goal around margins is to make sure that we continue to transform the portfolio to more and more of our IP. However, within these categories of storage, so it's easy to say block file and object, but there's a broad swath of use cases underneath those. And we just brought out the Alletra X10-K, which is based on the same hardware platform, it's an Alletra MPs, based on the same hardware platform with our object software. And one would ask, well, what are you doing with the other third party object offerings and file offerings in your portfolio. And so what we are targeting is GenAI and some data protection use cases right now. And the other pieces of our portfolio will fill in the other use cases. And this kind of transformation of the portfolio over the next couple of years, should allow us to increase margins with more of our own IP. And -- but we also want to make sure that we continue to cover a broad swath of the market. So there will likely always be a combination of third-party and IP, but really it's the balance within the storage portfolio that we're working to manage.

Got it. And just how do you see the opportunity for HPE and kind of that commodity flash or QLC market? And -- so when do you expect to kind of have a product there?

We already have a product. So in Alletra MP last fall, we introduced QLC drives and they will be in the X10-K this -- in the next month or so. And so here is the opportunity for us. So QLC, slightly lower performance, bigger drives, ideally suited in objects. But in the block side, where they're ideally suited is in customers who had hybrid arrays before. So they might have had a combination of hard drives and flash, and QLC offers the economics. So now you can replace. You can replace a hybrid array like some of our older arrays. And so you've got a nice upgrade path for customers. And the benefit they get is they get more performance, they get better TCO because flash, and this is, I think, what we tend to forget flash is greener, uses less power, and therefore, you have smaller footprints, bigger drives. And so there's a whole slew of benefits in going from a hybrid array to a QLC-based array. And as I said, we introduced this last fall, so.

Sorry, I'm just talking to myself without mute on. So maybe moving on just kind of to the overall business, kind of wrapping up. You guys mentioned the $300 million [ RIF ] on the last earnings call. And just how does that kind of impact your business? And where does that we view most focused going forward?

So the -- I mean we are taking a number of cost management actions. And we're trying to be incredibly surgical about this making sure that we continue to put our resources behind the parts of our portfolio and our strategy that are critical to us going forward. And so within my business, I think actually the benefit for us is there are areas where we can use technologies like GenAI to help us optimize. And so we kind of have some tools in our tool basket to help us manage costs and also in some ways, increased productivity. And then we have a much more focused portfolio than we did 1.5 years ago. And so that is also helping us hit our cost management goals. So we're trying to stay focused on the strategy while making sure that we do cost containment.

Got it. And a question came in on the line and Shannon Cross, who's Chief Strategy Officer, may want to answer this as well. But just kind of any shifts to either the GreenLake business or to the business at large, just given all of the tariff uncertainty.

Yes, I appreciate. I'll take that one. I think -- I'm not going to speak specifically to GreenLake, but I think, look, it's a fluid situation. I can point you back to the comments we made on the call, the earnings call about the impact that we were expecting to our earnings, which was $0.04 this quarter, [ $1.80 ] as you go through the year. I think we have a global footprint, the majority of our -- over 50% of our revenue, about 60% is generated outside the U.S. And I'd also point out that I think it's been pretty widely reported that servers and storage are -- they are made in Mexico or USMCA compliant. So we're working through it, and we remain optimistic regarding how we can manage the situation as well.

Got it. Shannon, I appreciate you coming on for the difficult question. Fidelma, I'm going to kind of wrap up with you, you're also CTO, if we could just kind of get your perspective on emerging technologies and trends you see in the market and just kind of what excites you kind of going forward?

Yes. It's -- what's really exciting is actually I've been in the business for quite a while, and infrastructure is cool again. And so we've spent many, many years where we didn't talk about infrastructure. So just from a personal perspective, the fact that we now are talking about massive changes in infrastructure. And in some ways, new operating systems with GenAI is just -- it's just very exciting. And so obviously, GenAI is top of the list. I think the second piece is the sovereignty and the rise of kind of the deglobalization in some ways. That is also a large trend because what we have to do as an industry is really rethink what does that mean? We're going in some ways from a centralized world, back to a decentralized world, which will also drive innovation in the tech industry. The tech industry has proven that with all these changes out of that comes opportunity. The other one is the continued importance of networking everywhere, everywhere. This is -- it's funny. We've got these 2 things going on. We have got deglobalization and networking. And everybody wants to be connected, but we don't want to be all connected together. And so on networking, the transformation that AI is going to make on networking, the importance it's going to drive and the innovations that that's going to find. So if you if you look at it and you say, even if you take a very simple case, I'm a networking administrator, okay, critical positions in most companies. The network goes down or something goes down and the network administrator gets blamed all the time. And of course, there's massive panic within the company because nobody can log on. Well, today, what happens is you try to figure out what's wrong, but it's very difficult. And you also can't figure out if it was the network, if it was the main application. And so now there's a -- it's going to come very soon as digital twins for network management. And so you build a digital twin of all of your network. And so now you can simulate issues. You can simulate fixes. So instead of going and trying to do a fix and it fails again, and everybody is still at you, you now do this within the digital twins. I think you're going to see an emergence across operations of all of these digital twins, really driven by GenAI and all of the use cases there. And then finally, it wouldn't be good to talk about anything if we didn't talk about quantum. And within HPE Labs, we've been working on a number of quantum projects for the last number of years. I think you're going to see quantum co-processing with GenAI and so as we emerge through the back half of this century -- of the century, of the decade, maybe it's because this year has felt so long. But as we merge through the back half of this, I think, you're going to see -- I mean you saw the Microsoft announcement, there's a lot of other work going on across the industry. You're going to see, in certain cases, quantum co-processing with GenAI. And so that's what I see going forward. It's a great time to be in the tech industry and infrastructure.

All right. Perfect. Well, Fidelma, this has been a super helpful conversation, and it's been -- taught us a lot about kind of an exciting business within HPE. So thank you for taking the time today.

Thanks, Meta, and lovely to meet you, and thank you to everybody on the webcast for joining in.

Great. Any questions, just follow up with the HPE IR team, and they'd be happy to chat. Thanks.

Thank you for attending today's webcast. The webcast has now concluded, and you may now disconnect your webcast line, and have a good day.