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

Welcome back to the Jefferies software conference. We're really happy to have with us the CTO and Head of Software, Kumar, at Hewlett Packard. This is a unique opportunity to learn more about what HP is up to on their software business. Kumar joined through the acquisition of BlueData, where he was the co-founder and CEO. Prior to that time, he was VP of R&D at VMware. I'm sure many of you will ask questions later about the transition with VMware today. We just had VMware on. But Kumar, thank you so much for joining. Marcus is also on. Marcus had to go through a couple of quick housekeeping items, and then we'll go through a quick presentation from Kumar.

Super. Thanks, Brent. Before we start, let me take a moment to read our disclosures. You will hear some forward-looking statements in today's discussion. These are based on risks and assumptions that are described in our annual report on Form 10-K and Form 10-Q. Our actual results could differ materially, and we assume no obligation to update. More details can be found on our website, investors.hpe.com, and our recent earnings announcement dated December 1. So with that, let me turn it back to you, Brent.

Great. Thank you. Kumar, why don't you kick us off? Bring us through the story, and we'll have some questions at the end from the audience and from our team. Thanks again for joining.

Thanks, Brad, and hello, everyone, and good afternoon from the West Coast. Again, my name is Kumar Sreekanti. I'm the CTO and Head of Software Business at HPE. Maybe just probably taking a little bit of more detailed introduction. Thanks to Brent that he gave a good introduction. At HPE, I have 2 roles. As a CTO, I drive the corporate strategy. I advise Antonio on various technical issues, we look -- when we look at M&A and Pathfinder investments, we look at the technology. Also, I have a responsibility for the software business, which I think mostly I'll focus today on the presentation, how the software business is integrated into HPE, how it drives the growth for the Ezmeral product. We are obviously looking at how do the -- rich software platform for HPE and importing the customers to adapt the containers technology. And I'll also speak in quite detail about the container technology that we're doing today. As a background, we were -- I was a co-founder and CEO of BlueData. BlueData was acquired or bought approximately 2 years ago by HPE. BlueData was very focused on 2 things: container technology transition and the AI, ML and big data workloads. Even though the technology that we've built is very broad or -- as a start-up company, we're always focused. And so we did -- we built many, many technologies that are very specific to AI/ML, and we'll talk about it in my presentation here. Prior to founding BlueData, I was a VP of R&D at VMware. I ran most of the storage, networking and clustering technologies. vSAN was one of the technologies that I initially brought to the R&D. Prior to VMware, I've been in many technology companies, including as a VP of R&D at Akamai. So let me start with -- you probably have seen this slide. This is what Antonio and team has presented at the SAM. I thought it will be a good point to bring is customers, what we see is -- what cloud has done for the enterprise customers is they want the same cloud-like experience on on-prem as well. They want speed. They want flexibility. They want cost controls. They want to have an edge to the cloud strategy. Obviously, they would like to see the cloud-like experience all across the board. The one interesting thing that it's probably worth noting is what we feel all the low-hanging fruit, which we call nonstateful application for the most part, has transitioned to the cloud. By one -- there's a couple of estimates. For example, IDC estimates about 70% -- IDC and Gartner, 70% of the workloads are still on-prem. Many of you may have heard at the recent re:Invent, AWS has said, something in the order of 90-plus percent workloads are still on-prem. We believe that data gravity, security, latency, performance and compliance issues are predominantly helping -- or not helping for the transition. Data gravity is a very complex issue. One of the -- we actually talked about at the SAM, it's about a $22 billion market that we think for as a Service for software and hardware on-prem by about 2023. One of the things that really excites me is this kind of discontinuities in the marketplace, which is the containers, which is now, I think, synonymous with application modernization, which is synonymous with how do I make sure I transition into the post-COVID world. These are once-in-a-decade kind of opportunity for all of us -- for us as a company, and we want to help our customers make that transition. From HPE, how do I think and how do I help my peers and my boss to think about what are the 3 key pillars of our strategy? And I thought I would start with that. Even though I'm going to spend all the -- most of the time on HPE Ezmeral software. But internally, here is how we think about it. Our North Star is HPE GreenLake as a Service. To get to the North Star, we see 3 pillars, which is obviously HPE Ezmeral software, which we have announced the software platform at Discover last year. And then the second one is HPE is a very advanced leader in the hardware. How do you provide that software and hardware integrated to provide the experience, the workload optimized, providing them the cost effectiveness, in many applications, you can run on a bare metal. You don't have to pay the tax, sometimes you have to pay significant tax for virtualization. And how do you get the high performance -- highest performance per workload and vertically integrate. The days of taking a piece of hardware and software and putting them together is over. Really, the customers would like to have a business solution. And then, of course, how can we consume this in the marketplace, on-premise, managed by you, by HPE, in a self-service manner, and you pay for what you use for. So those are the 2 and 3 pillars. So let me come back to the Ezmeral, right? Ezmeral is -- it's a technology that's actually very focused, purpose-built on the -- for the containers. It provides cloud-like experience on-premise. It's obviously based on open-source Kubernetes. And it's -- one of the unique technologies, as I said, that BlueData has built is very data-centric applications. We believe the majority of the new workload starts to go in that direction, and you are seeing as well, is that AI is the new -- there is a famous thing like software eating the world 10 years ago, Marc Andreessen said. I have a new one, software is re-eating the world artificially, intelligently this time, right? And then we -- HPE is a unique company that we have edge-to-cloud, scale-out data fabric technology. This is another technology that, after I became CTO, we have acquired through MapR. So these are -- the beauty here is you can run the HPE Ezmeral software on-prem, on any hardware, not only HPE hardware, on cloud or on a hybrid. And you can actually also run it on GreenLake, and we can manage it for you. So it's -- we think the world is hybrid, and I think everybody now agree the world is hybrid. So this is a classic fit into that hybrid world. A bit more detail into this -- our platform. One of the things that quite often is missing is when we built these technologies, both at the BlueData and MapR, and then now we got all the HPE, both organic and inorganic, we brought all this together, how do you run both cloud -- noncloud-native applications and cloud-native applications. And just a little bit of a primer, cloud-native application meaning it's kind of synonymous with containerization, microservices, but data legacy, many, many legacy applications. When we go talk to the customers, many customers say, we have this massive amount of existing applications. They would like to modernize them, but they can't refactor overnight. So this is a unique platform that we built that allows you to run on the same platform, with the same single pane of glass that you can run both your cloud-native applications and noncloud-native applications. What is unique about this platform? It's obviously self-service. It's -- if you go through the GreenLake, you do the pay per use, we can actually run both manage for you or you can manage. And it runs on the core, as I said, on-prem, on the edge, on the public cloud as well. What is the uniqueness about the HPE Ezmeral container software? And what is the differentiated IP, right? So when we built these technologies, as I said, some of the core components are from the -- when we did the BlueData is that how do we get specific to the new workloads. And those are the workloads which we call big data analytics and ML ops. This is a unique platform. You can run H2O, Jupyter Notebooks. You can run Spark. You can run Kafka applications. You can also run traditional applications. And it is based on a Kubernetes open source, so there is no lock-in and yet run top of the trunk open source. And it has unique data fabric technologies with the data management that is built into it. So typically, when you run stateful applications, they tend to have a need for a persistent store because a stateful application implies if it fails, it has to come back, and then it has to know what state it was in. And then for which, it needs some kind of a persistency. So we have actually tagged all of them together in this platform so the customers can have a -- both infrastructure services, they are enterprise secure, they are enterprise grade, they run open source, they give you an application platform; and then they have data services. One unique capability that we have built in, we use the term DataTap. It's a kind of an interesting extension to data lake. So if you build some data lakes with some, let's say, older technologies, so you don't want to move the data, and we have now come across many customers, and they try to move these large amounts of data, and the cost of either moving or egress costs, ingress costs are so heavy. So we have actually built technology that you can actually leave the data there, and you build these new platforms. And these new platforms can access this data that you have on your on-prem, and you can use them to run your analytics on them without having to copy the data. Other problem with the data is generally when you tend to copy the data, you have a source-of-truth problem, meaning you'll end up with multiple copies. So with the DataTap technologies, you have a very unique capability. Other thing that we do is we have built an ML ops platform. This is one of the problems that you see. So all the newer applications have not only their -- containerized their modern way of implementing. But they also have a component of -- there's a large amount of data generated. I don't think I need to say this. And then -- but generating a large amount of data does not give me anything unless I find insights from that data, meaning I have to be able to run some analysis on it. Earlier, it used to be analytics. Now it is AI application. But the problem with the AI application is not just developing the model, it's developing the model, training the model, deploying the model and retraining the model. So this is ML ops platform that we have built. So this is a turnkey solution. This is not just only CUDA. This is another great application. There's no lock-in here. You actually run any applications that you build. So typically, what I hear from customers is when they go to a specific public cloud, they're worried about the walled garden. So here, you can run all open source applications, Jupyter Notebooks are open, but you actually have a platform that gives you the ability to do that. Another application that I'll talk about, you may have heard in our SAM thing, is Splunk is one of the very big applications that we have seen being deployed. Now they'd like to have a cloud-like solution but want to use Splunk on-prem. And that's another great application. There are other ones like DataRobot, and I talked about the H2O, TensorFlow. One last thing is graphic processors tend to be very efficient, but they are very expensive. So with this platform, you can actually virtualize them, and you can run them for multiple applications. So you don't have to dedicate it for the customers. On this slide, I wanted just to give you what's the difference and why this platform is unique. So first and foremost is it's purpose-built from the ground up. It's not a platform that's trying to protect the legacy of -- legacy platforms, like, let's say, most of you are familiar with innovator's dilemma, right? Whenever there is a legacy and I have a lot of business coming out of it, I don't want to -- new applications to take away, so I want to protect it. So there is no such thing here. This is built up from the ground with focus particularly on AI/ML applications, even though it actually allows you to run any application you like. But it has unique capabilities and unique features that are meant for running AI/ML applications. Enterprise grade, obviously, we curate this -- we curate whatever the Kubernetes source we put out. It's been tested on hundreds and thousands of nodes. It's a hybrid. As I said, the world is hybrid today, right? You can't just say, I want to run on a cloud, and I can't bring it home, then I will repurpose the application. Now you can run the same platform anywhere you like. It's -- another very key feature, as I said, it's data fabric and underlying phenomenal technology. And I'll talk about in a second the history of the technology, which came from MapR acquisition that's pre-integrated in this platform. So you as a customer, you don't have to worry about it. Many of you know, MapR was 1 of the 3 key technologies back in the early 2010. MapR runs some of the largest databases, including the India's Aadhaar card. I think it's about 1.5 billion records. And that technology is now available via HPE Ezmeral. We can also -- as I talked about DataTap, it connects to the external data. And last but not least, it's -- a very interesting evolution that we are going to see is when the containers become more mainstream, they have no necessity to run in a virtualized fashion. Actually, running containers on virtualization is it somewhat double virtualizing it, right? Of course, the argument against this, containers are not mature enough, and virtual merchants are more secure so we'd like to run them. And that is true. As an engineer, I can tell you, early days of containers, if you're running an enterprise application, you won't want to be in. But containers have matured significantly. There are more container store security advancements I have seen in the last 4 or 5 years. Containers also have solved the networking problems inside. They have also solved the storage interface problems. So now containers can run bare metal that gives you the -- if you saw my second pillar, that gives you the best performance for the underlying bare metal. So we can actually run containers in bare metal. I'm very sure many of you had a lot of questions. So I will try to finish in a couple of minutes here, and then we can answer the questions. So we -- Ezmeral, formally announced early in summer last year, we have talked about Wells Fargo at our SAM. We also have customers like BMW, Barclays, GM Financial and other customers. So one question in all of your minds is what is that these customers are doing, and we'll talk about Wells Fargo in a second. But typically, what we have seen is, I've said this multiple times, I think there's a lot of data. People like to harness the information out of the data. And they are actually seeing that harnessing those applications come in 2 categories, either you're writing the new applications that tend to be microservices and containerized. And there are also older applications that they would like to drop into the containers. And then you can run in a containerized fashion. So that way, you are not dedicating the hardware. You utilize the hardware in a much better fashion. You can also run in the cloud. And what we have seen is, typically, there's edge data. BMW is a very good example of a lot of autonomous driving edge data, and how do you process it, how do you create it, how do you run disconnected operations and things like that. And the last one, we wanted -- I wanted to share with you, this is a very unique opportunity here that I thought would be interesting. So this is an application that they looked at all the container platforms and -- on the cloud. And they clearly decide, after running through a significant amount of benchmarking, it came through, how do you serve a specific application, in this case, Splunk, in a scalable platform, in a cloud-like experience, on-prem, at a cost advantage that's at a distinct -- at a significant cost advantage and then modernize their application without having to rewrite multiple applications. So we are very pleased to have this, and we hope, next time I talk to you, we'll have 10 of these or 15 of these. With that, I'm going to pause and see if there are questions, and we can have a dialogue. Thank you, Brent.

Thank you, Kumar. Appreciate the overview. [Operator Instructions] There is a question. Can you just talk about the differentiation of how you're going after this versus kind of what IBM and Red Hat are doing?

Yes. So there's -- I would say there are probably 3 different ones, right? HPE is the most trusted enterprise partner that I have come across, right? So one of -- we have a very large services organization. Many of these problems are extremely difficult for the customers. The days are gone where you can drop pieces of software on top of hardware and the next day it works, right? So you need some help and solution base, so with our ability for our Pointnext and A&PS services, with our hardware technologies and our trust of our customer. And if you look at those 3 triangles and put them on top of GreenLake, so customers would like to have somebody help run them on-prem, just get to cloud-like experience, right? I think -- I always joke about this. What is the cloud's effect on us is everybody wants like a cloud-like experience and not going to cloud. So that's what our focus is on, how do we help the customers to modernize and have the data where it is. So I talk about bringing compute to the data as opposed to taking data to compute. So that's the -- and then the -- as I said, we have some unique capabilities around AI/ML and the big data. So that's -- we naturally see the customers gravitate into that.

That's great. Within the large enterprise installed base, how far along are customers moving the existing apps to on-prem private clouds? And what are the drivers, catalysts?

It's very early, Brent and team, right? But I do say COVID has changed -- and I think Satya said this, right, we saw 2 years of advancement in 2 months because everybody is starting to realize they have to modernize their applications in a hurry. And then now COVID took the time factor out of it. So they are moving very quickly, but it's very early days, right? The hindrance for them is the data gravity. And every CIO and CTO that I've talked to is -- has these massive amounts of data, and they have security and compliance and latency issues. So -- but they would like their apps to be modernized. So what we are -- as I said often, quite often, what we are bringing them is you can modernize applications while you are running your existing applications on the container platform. If I just may indulge on Wells Fargo a little bit more here. They want a radical change in the costs. They want a cloud-like experience but not by giving an arm and leg to the public cloud. And they want to have a better performance. And then at the end, for us, for HPE software, it's a TAM expanding for us, right? I mean we are getting the new workloads on-prem.

When you think about the company not necessarily having amazing success selling stand-alone software in the past, what changes this time? What's different now with the software strategy that didn't exist before?

I would like to believe it's me. No. No.

Other than you. That's the obvious choice. But...

Yes. I know it's -- I think it's the -- what Micro Focus and what -- they were on a different platform. I believe HPE's sweet spot is IaaS and PaaS. And because we build the best -- we -- I mean, at the end of the day, we supply 1/3 of the servers in the world, and we supply a significant amount of storage. We supply a significant amount of supercomputing. And we are -- we supply a significant amount of networking gear. So IaaS and PaaS is a very natural extension for us. This is also an opportunity for us, the dichotomy, the discontinuity in the industry going from the virtualization -- so we went from client server computing to the cloud, to the virtualization. So this is the next modernization. So we are at the unique opportunity. And I call this, it's a unique strategy for a unique company at a unique time. And there is no established players in here. There's nobody established. Nobody has an 85% market share around containers. So -- and then we have a technology that makes a difference for them. So that's why I think all these 3 pieces come together for us. And I think it's the opportunity to help the customers to modernize on the containers and show them that HPE is a great partner.

That's a pretty staggering figure. 1/3 of the servers in the world gives you a really great opportunity. When you think about the commitment you're getting from the team and their commitment to software, can you just talk to that and how this sits relative to the overall priorities? I think many have asked, hey, like it seems logical that you would double down. We're all software people. We're not hardware people. But it just seems like it's -- this is kind of a natural double down, if you will. And when you think about the resources you're getting, distribution, sales, can you just talk to the commitment that you're seeing from the team to make this move?

Yes. And Antonio asked me to take this role exactly for that reason, right? So I not only have -- I have a job -- I have a day job and a night job, not only as CTO but as a head of the software. And the point is exactly for that. It is software is the glue to bring all these technologies together. So if you saw my 3-pillar slide, that was actually underlying -- intent of me showing the slide is it's actually bringing all these technologies together. And then the second is that we have a very large worldwide sales force that is extremely experienced with enterprise, and all the sales forces now has Ezmeral in their bag, and they help our customers. So I believe I think we are -- and then we talked about Pointnext and A&PS. So between the GreenLake and our service capability, our sales capability, I think Antonio has made it very clear even at the SAM, I think this is a very key part of our strategy going forward.

That's great. And just from when you think about this re-architecture that clients are going through, can you just talk about the conversations that you're having now? It seems like these re-architecture initiations are really accelerating, and it seems like a great opportunity this year. Maybe just compare and contrast what you're seeing in the overall demand environment now versus what you saw last year.

Yes. I think you heard this from Antonio, I think I mentioned Satya, I think everybody, COVID has taken the time factor out, right? People are always thinking the best way to be -- you must have heard this, and I will say again, right? Tesla is not a car-making company. It's a software company that makes cars. Uber is not a rental company. It's a software company that lets you rent rides. So software has become the key element for all the enterprises in their business segment to get advances. And for them to get advanced, they have to rethink about how they collect the data, how they analyze the data and how the data helps them to make the decisions, right? So -- and COVID actually put an extra -- I call it this decade is going to be a decade of disruptions. I think every business has to disrupt themselves. Otherwise, it will be like Airbnb disrupting hotel industry and Uber disrupting taxi industry and Tesla disrupting other automobile industry. I think what they have to do is, for them to avoid getting disrupted, they have to advance and figure out how to do what they do better, at a better cost, at a faster rate and more intelligently. So every conversation we see has 3 components to it. One is the data. We are collecting more and more in data. And we want -- how do we make sure that we don't pay for arm and leg and how do we make sure -- second is how do I harness this data, and number three is I have to keep my costs down. And everybody is told that AI will help them to keep their costs down. So we are at the juncture of all these 3: the data, analyzing applications and using AI.

What's the best hidden gem that we can't see that you see every day? What's most exciting to you that maybe we just can't see as outsiders?

From HPE you mean?

Yes.

I think HPE is a great company. In fact, when I went to school in IIT, I always tell that we used to look up to the Valley and HPE and IBM and Intel. There's a lot of great things within HPE. Obviously, in the last few years, it has been somewhat tough. But I think my goal, working with Antonio and the leadership team, is put HPE back as the technology leader. There's a lot of culture and a great team.

That's great. Well, we really appreciate you sharing your perspective on the industry. Thanks so much for joining us this afternoon, and looking forward to staying in touch and learning more. So thanks again.

Thanks, Brent. Take care.