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

Conference. I'm Wamsi Mohan. I cover IT Hardware. Delighted to welcome HPE CFO, Tarek Robbiati. You all know Tarek. He's been CFO now since 2018. So seen some cycles here recently and navigated a lot of cross current. So Tarek, thanks so much for joining us here today.

Thank you, Wamsi, for having me one more time. Very happy to be here with you and your clients.

Thank you so much. So Tarek, it's been an interesting year, right, at SAM. I think you guided revenues 2% to 4%, then you went through a quarter, you kind of doubled to 5% to 7%. And more recently, it's come down on a reported basis to 1% to 3%. So can you talk a little bit about some of the dynamics that caused these kind of large swings in revenue?

Yes. Thank you, and it's a great way to start our dialogue this morning. If we put back things in the context of SAM 2022, which was in October of '22, at the time for us, there were a couple of major unknowns like for many other companies. One was the macroeconomic environment, but also and more so, how does that translate into foreign exchange risks. And we did anticipate a pretty substantial volatility in FX rates when we guided. And we did guide to the levels that you are -- you have mentioned. We had an assumption on FX risk that has fluctuated quite substantially in the first 6 months of this year so far. And so we had to take that into account in issuing new guidance for the full year-end. And so we came to the 1% to 3% full year revenue growth guidance in U.S. dollars, having to assume a 250 to 300 basis point headwind from FX rates. And so when you do the math and you add that back on top of the 1% to 3%, that is very much where we thought we would be, and that gives you the impact of foreign exchange rates on our business because more than 50% of our revenue for people in the audience who may not know comes from non-U.S. dollar countries, right? So that is a very important dynamic. The second dynamic that has manifested itself during the course of the year is, first, easing of all the supply constraint that we had been witnessing for the past 2 years. You may remember, '21, '22, the most spoken word was supply chain. No one speaks about supply chain in '23 anymore. It's no longer used. It may be out of fashion, which is not a bad thing. Before people didn't know what supply chain was. So supply constraints have eased dramatically. And this has to do with the pandemic coming to an end, factories resuming production in Asia, in particular. And therefore, when we exited fiscal year '22, we had an order book that was very, very elevated that we could not fulfill because we did not have the supply. And the pace at which the supply constraints were resolved has surprised us, quite frankly. And this is why we posted a very solid and very strong result in the first quarter of fiscal year '23. And when you have the supply and you want to make sure that you protect your order book and you have to use the supply to deliver to your customers, we did not want to see cancellations happening on that order book, and they haven't. And that explains to you the result in Q4, end of Q4 and certainly the result in Q1 of fiscal year '23. So the supply dynamics have changed. And then the third dynamic that we've witnessed in the first half of our fiscal year is very specific to the banking sector worldwide. I mean you are all fully aware of what has happened to 2 banks in the United States and a large bank in Switzerland during the course of the first quarter. And our business's largest customers are banks globally. We have a very substantial customer base in banks, manufacturing, telcos. These are the main industry verticals. And surely, you probably had a few e-mails on your end coming your way, Wamsi, this is what's going on. We've got to cut cost and so on. And IT departments and banks are certainly the first place where that cost pressure comes in and where people are asked to sweat assets for longer. And this explains to you why demand in Compute sequentially between Q1 and Q2 of our fiscal year '23 has declined in low single digits, as I've told you of my CEO, said on the earnings call 10 days ago. So these are the 3 dynamics. So in summary, FX; second, supply chain -- FX is a headwind, tailwind. It always is a little bit difficult to predict. Supply chain is a massive tailwind and then demand in some sectors has waned during the course of Q2 '23.

Okay. Yes, that's a helpful bridge. I appreciate you going through that. In the quarter, you saw very standout growth in Intelligent Edge. And we've gotten a lot of questions trying to figure out how sustainable some of this is in light of the fact that we did see outsize growth in other segments because of backlog reasons and Edge had big backlog too. So can you just talk about some of the dynamics that transpired? And if you look beyond just sort of this fiscal year, like how should we be thinking about gross margin in Intelligent Edge?

So our Intelligent Edge -- we're extremely pleased with the performance of our Intelligent Edge business, and we've been prepping that business unit for some time, and it's nice to see the fruits of those results. Back then in October '22 at our Securities Analyst Meeting, I said this is a Rule of 40 business. Not many people believed me. But yes, here we are. We are no longer at Rule of 40. We're at rule of 80 if you look at the latest result, which is obviously exceptional. And the way we have to think about the Intelligent Edge market is that we do see durable growth on a mid-single-digit basis for the market, and Aruba being a share taker within that. Having said that, when you take the lens of 3- to 4-year forward, that growth of mid-single digits ebbs and flows, right? So there may be a moment of pause that takes place in the short term, but there is durable demand for networking on a mid-single-digit basis if you take a long-term view. And why is that? It's because of a couple of factors. Well, first of all, the vast majority of the data that is generating in the world is generated at the Edge, meaning at the fringes of the enterprise. And second, in any IT infrastructure, the bottleneck is always the network. The thing that keeps you from not going as fast is the network. And therefore, there's always demand to upgrade technology and networking. There's always demand to do things better. And we feel very comfortable with our business Aruba and what they were able to do. This business has been performing exceptionally well, thanks to the Aruba Cloud platform, which is now powering GreenLake. And that gives a handle to chief officers on the infrastructure that was unprecedented. It's a control plane that allows you to very quickly understand how the network is performing because it's a critical bottleneck, isolated areas that may not be performing well, fixing things that need to be fixing and so on. So I'll give you one example. It's a real-life example of a large American institution who had a problem connecting iPads to their network that is powered by Aruba. So obviously, they called us and they say, "What's going on?" So we dig into it and figured out [Technical Difficulty] So that's a complicated problem to solve because you have to recall every iPad if you want to go and change the configuration in the iPad. And people have daily jobs and they use those iPads for their jobs. So it's not practical to ask for a recall. And also this institution has the problem for their employee network, but they also have a sort of proprietary network available for their own customers that people can connect to also with the iPad. So needless to say, it's also a customer problem for our customer, right? And so the Aruba team went on through the Aruba Central platform, figured out what was going on and rather than recalling iPads, they pushed out a patch in this very large institution with several hundreds of points in the country. Problem gone in a matter of days. Right? That's not easy to do unless you have a very, very intelligent capability, both in the central point but also your endpoint are capable of dealing with the issue. So the endpoint have to be sufficiently smart to take on the patches and reset and so on and so forth. And needless to say, we've learned from that because, obviously, if such a problem was manifesting itself in that particular institution, chances were that could have happened elsewhere. And so Aruba Central allowed the Aruba team to push the patch everywhere. So that is something that essentially customers are very, very much resonating with. They understand that power, try and do that with the solutions from our competitors. Not possible. In some cases, it takes months or quarters or years. And that is helping our customers understand that they need to manage their total cost of ownership of an infrastructure -- it is so critical to their own business in a very, very efficient way. So that explains to you the success of Aruba. The other thing we're doing with Aruba is we are expanding the scope of the portfolio horizontally. And this is a very important strategic decision we made because where Aruba started was obviously with Wi-Fi and then campus switching and so on has been the stellar success of it. But beyond that now, we are investing in security. This is why we acquired a company called SASE -- in the SASE space called Axis, pardon me. Axis is a fantastic platform that provides zero trust networking capabilities to assure the security of a network. That is combining extremely well with Silver Peak, which we acquired back in 2020, Wamsi. And because we know that when people look at their network connectivity with SD-WAN, they also have to look at security. So we estimate that for every dollar of SD-WAN being sold, there is a dollar of security to be sold with it, and that's why we acquired Axis, which is a very important vector of growth for Aruba in the future. And then the other area where we're spending a lot of time -- a lot of time on is data center networking. Aruba's data center networking business is very interesting. It's growing really fast. It is powered by the same operating system that powers the campus switches, which gives even this greater control to the network officers inside companies that operating system is called AOS, ArubaOS, and it's doing extremely well. And then the third vector of growth there for us in that we're adding is private 5G. We did a small acquisition called -- of a company called Athonet which is an outstanding company in the space that has unique capabilities to cater for very demanding customers such as ministries of defenses around the world. And that business transaction is about to close. And at some point, this will be part of Aruba. Sorry for the lengthy answer, but I thought it would be important for me to give you color once.

It's helpful context. What about on the margin side, do you think that that sort of mid-20s margin is something that's sustainable?

Yes. I think Aruba's margins are -- very comfortable with the margin profiles in the 20s. Obviously -- to the extent that you keep the OpEx under control. If there is high growth, you have higher margins. There is lower growth, there is slightly lower margins. But we feel that from an OpEx structure standpoint, Aruba is extremely well handled. If anything, we're tweaking the mix between how many dollars we apportioned to R&D versus sales and that there's always a bouncing act, but the cost structure of Aruba is perfectly well suited for the space.

And when you think about talking about this as a Rule of 40 business, is there implied in there how many points of share gain in the space is implied kind of in the thought process of getting to a Rule of 40 business?

Market share is always an output. What really matters to us is whether we can grow the revenue in the way we like in the segments of the market that we are pursuing. Campus networking with Wi-Fi, which is data center networking, SASE and SD-WAN and also private 5G and we feel that Aruba now has a very, very broad portfolio, possibly one of the best portfolios in the industry. And so we do expect growth there, but the industry is also growing. And so if you really look at Cisco, for example, grew. Look at Juniper, they grew. Extreme, they grew. So it's a very lively space. And what we need to do is to stay focused and continue to grow. There is a market share ambition, but there is no point for us to talk about it. Now what we need to do is to stay focused on executing.

Okay. Maybe pivoting quickly to H3C, you announced recently about your $3.5 billion value that you're going to get out of the H3C JV. So a couple of things. What are the other milestones that need to clear? And what is the timing that you anticipate getting the cash? Can you talk about how much that might get taxed at? And what would you do with the proceeds?

Okay. There's a lot of questions in here. So first of all, very, very, very pleased to have reached this agreement on our stake and the value of our stake in H3C with our partners of Unis. It was very important that we got to an agreement because the alternative would have been a lot more complicated, and would have involved arbitration proceedings, which are lengthy, random and never assured. So now we have a very clear price that the 2 parties totally agree on. Their board approved the transaction. And now it's a matter of passing a number of regulatory hurdles. And the regulatory hurdles are only in China. There is no approval required from any U.S. regulatory agencies. That's important. And it is in the normal course of business. These approvals were contemplated at the time that the joint venture was conceived. And so when you really think about this, this is very much the ordinary course of business, and it's an important set of approvals that will give Unis full control over H3C, which is the most important and most valuable asset that they have in their overall holdings. So this takes -- it's a process that takes about 6 to 12 months to go through the administrative hurdles. And -- but we don't see any reasons why this wouldn't end successfully. So we are pretty confident that it will. With respect to the proceeds, the $3.5 billion figure is gross. I would not venture giving you a view on tax because we're working through that. And if you really think about it, it goes full circle because the approvals to pay us $3.5 billion involve agencies from the Chinese government who is also going to tax us on the way out. And so we'll have to find a way to obviously manage that and maximize the net proceeds. So we'll just have to do what we normally do from a tax planning standpoint, and that's it. How are we going to use the proceeds? Well, we'll have to think about our capital management policy one more time, which is always a balancing act between capital returns to shareholders and investments for future growth. And we are going to have to deconsolidate H3C once the sale truly materializes, right? And H3C contributes a fair bit of EPS each year in cash flow because they were one of the few companies that was able to truly extract cash dividends out of China, right? And so that is something we'll have to address and we need to continue to grow our EPS and grow our cash flow. So it's part of the balancing act involves that reflection of if we make an investment, what do we invest in to make up for that shortfall that will come upon deconsolidating H3C. I hope I've answered your questions.

Yes. I guess, timing. Is there anything on sort of when you think this approval process will complete?

6 to 12 months.

Okay. And then lastly, on the proceeds itself, right, would you consider that to be part of what you would define as cash flow -- free cash flow, the proceeds from this entity when we think about talking about like future free cash flows of this year, we're talking about $1.9 billion...

No, it's not included in our free cash flow definition, right? No, it's not part of the free cash flow guide that we gave you. That's separate.

Right. Thanks for clarifying that. I guess I'd like to jump to maybe AI as a topic because it's obviously very topical, but you guys have had a segment HPC & AI before AI became a big thing here this year. But when you think about who are the customers that are driving demand over there, what are you seeing? And over the course of the next few years, how do you see the demand for AI on-premise? And what assets do you think or IP do you bring to the table for that?

So AI is unquestionably incredibly hot space ever since the 1st of January. And I think whatever happens to the AI sector is going to be very much a story that has been seen before by some aspects and unseen by other aspects. So it's -- let me start with the unseen portion, which is easier. I would say AI is going to change our world. There's no question about that, and it's going to change our world in multiple ways. I mean, you all heard about ChatGPT and what can be done through ChatGPT. And if you think about, for example, a software developer, who -- an average software developer, who has to solve an average software code problem. Typically, what that person does is that he or she connects to a forum like Stack Overflow to ask questions, how do I solve this problem? And they find a solution because someone there is generous enough to tell them how to go. Well, if you replace that with a machine, then you understand that the machine can give you the code routine itself, and there's not much else you need to do. And so there are jobs that will go and new jobs that will be created. And that's the new unique AI area -- what is going to bring to us. The movie that we've seen before in here is the one that I can tell you from my vantage point as CFO, having gone through the dot-com area, some of you are too young to remember that, but be aware of the hype, okay? And that is the thing that is right now the most important thing to remember. When these things are saying to be changing the world, there's a lot of hype. So we do see a lot of Bitcoin mining companies rebadging themselves as AI companies. And so this is more than anecdotal. So be careful with that, right? So then when you really want to separate the kids from the adults and we look at who's going to be building the real stuff, we believe that we become very relevant in that space. And the reason why we become very relevant in that space, it's because we have a very solid understanding of how you build an AI machine. That understanding comes from our experience in supercomputing. And the largest supercomputer in the world is Frontier. Frontier is a system that we rolled out in the early this calendar year in its final form for the Department of Energy. Anybody knows how many GPUs does Frontier carry? 10,000? 37,000 is the number, right? So it's an enormous machine, right? And it's -- if you really look at its power, it's 1.2 exabytes of power, which makes it the most powerful supercomputer taken -- even if you take the next 7 put together, you don't get that power, okay? So -- and why is that important? Why is that important? It's because the quality of your AI model is dependent on the data set that you've tested that model with. And that data set, the bigger it is, the better your model, the smarter your model becomes. And you need an enormous amount of data to do that. And we break the market into 3 spaces, large language models, generative AI, training of these models on smaller machines and inference at the Edge. And so we see opportunities across this whole chain. And this is why, Wamsi, I have taken the long answer to tell you, there will be lots of opportunities on-prem. Why? It's because you need an enormous amount of data to be processed and therefore, that opens up tremendous opportunities for us in that space. And if you don't have the computational know-how to build a machine like this, here's what's happening. If one of those GPU fails or a few fail, the job fails. So your model doesn't work anymore, right? So you've got to have someone, who stands behind the machine, to really deliver that. And that's why we've had an unprecedented success in the quarter with $800 million orders in AI alone as we disclosed in our earnings release and counting. And we get a lot of orders here, and there is a lot of, as always, from the dot-com area, pick up the lesson a lot of money pouring into that space. So we have to be very careful and very selective in what orders we take and what orders we pass. And I spend a lot of time with my team to be extremely selective in that space because these are very capital-intensive businesses that require large investments. And then at that point, you actually think in terms of credit risk and asset risk and all the beautiful things that CFOs loves to talk about.

Maybe let's quickly pause here and see if there are any questions from the audience. One right here. Do you mind just waiting for the mic. All the way up front here, please.

Just one quick follow-up first on while we're on AI or kind of just high performance compute. That segment, the growth -- or the operating margins have declined over the last 2 or 3 years. And so what are you doing to turn that around? Because you just said you put one of the most powerful supercomputers out in the market, and yet it doesn't seem like you're really getting paid for it.

Yes. I mean it's a very valid question. The issue is that this is a very lumpy business because when the revenue comes in, if your cost structure doesn't change, your margin goes up. But to the extent that the revenue isn't there, you're going to have an impact on margins. And so what we need to do is to address that lumpiness. And the way we address the lumpiness is by changing the business model to become much more ratable, much more as-a-Service. And as part of the $800 million figure that we gave you, there is a very substantial GreenLake order in there from a very large cloud provider, right? And so we know how to build that stuff. We know how to think about the capital intensity, the asset risk and the credit risk because you're entering a space, which is the cloud space really, very much the cloud space, where you have to take into account an initial investment and how you recoup that investment over time with a ratable revenue from multiple -- single tenants or multiple tenants, depending on the situation. So that will take a little bit of time. But we have the view that this is a business that will generate durable margins moving forward, and we'll talk more about that first at Discover in -- at the end of this month and also in our SAM, Securities Analyst Meeting, in October '23.

Okay. Great. And then just one follow-up, just completely different on share buybacks. Last few years, you guys lately -- senior management ratio has been pretty [Technical Difficulty] but yet if I look at it, your stock buybacks is barely offsetting dilution. So your messaging is kind of different. If you believe your stock is undervalued, why are you not buying back more stock?

It's a great question, and that's a really good question, and thank you for asking it. So we -- you're correct. The buybacks are offsetting dilution and because dilution is important for our shareholders. But what you have also to reckon with is the cyclicality of our Compute business, which is representing 40% of the revenues of the company. And therefore -- and if you look at it on an operating profit basis, it represents a large portion of the operating profit of HPE. As this fluctuates and the fluctuations are really severe with commodity prices rising, commodity prices declining, then you have to reckon with volatility in free cash flow and, therefore, manage your capital accordingly. And what we are effecting is a very profound shift in the portfolio mix to avoid having that volatility. And that is why the message that we wanted to convey, Antonio and I, during the earnings release was so much about the portfolio mix shift. And we are -- if you really look at the trends and you can do your extrapolations and look at where Aruba is and where Compute is, they have very different dynamics, those 2 businesses, in the upcoming 18, 24 months. Aruba will represent -- already 19% of the revenue of the company and going to be above the 20% mark and a very large portion of the operating profit and growing. And Compute is on a different trajectory. It is 40% of the revenue and a large portion of the profits, but it is going to fluctuate. And the margins of Compute, like I said many, many times, are going to revert back to the 11%, 13% of OP like we always said in our long-term model. Today, on Compute, we are at 15.2% operating profit margin. No other player in the industry has extracted those margins. If you compare Compute performance to Dell or Lenovo, there's no comparison. No matter what value ascribed to the OP coming from our main competitor, Dell. So the portfolio mix is at the heart of it. And that -- once that portfolio mix changes, we may reconsider a different capital allocation policy. But for the meantime, I've got to be very, very diligent in the way we manage the capital management policy. The way we have with the balancing act that we always said has to be there, remunerating our shareholders and at the same time, investing to pull the company out of a cyclical business to become a much more sustainable, profitable growth business than before.

Thank you, Tarek. Unfortunately, we're all out of time. So we'll have to wrap it over there. I really appreciate you taking the time to be here today.

Well, thank you for having me. Hopefully, you all find it useful.