Everyone for joining. Welcome to day two of Bank of America's global tech conference. I'm Wamsi Mohan, IT Hardware Supply Chain Analyst here at the bank. It's my pleasure to welcome HPE to our conference today. We have EVP and CFO Marie Myers. Marie was prior CFO at HPQ, has had a long tenure at both the HPs and combined HP prior to that, and so knows all the assets really well. Delighted to have you over here, Marie. Thank you for joining us.
Thank you for the opportunity. It's great to be here.
Yeah, of course. Before we get started, I will just read out these disclaimers. Marie's remarks may contain forward-looking statements, so please refer to HPE's SEC filings, including their most recent Form 10-Q for a discussion of the risk factors that relate to their business. With that out of the way, Marie, welcome. Maybe to kick it off, right? This is obviously somewhat topical, on everyone's minds. The Juniper deal, can you just update us on where things stand? And maybe just what's plan B if the deal doesn't go through?
Yeah, no, thanks. Good morning, everybody. It's a pleasure to be here. I would say on the Juniper transaction, it's pretty straightforward at this point in time. The litigation date is scheduled for July 9th. It's going to take a few days. The expectation is that the judge will take his time to make a decision post that period. We'll know the outcome of the Juniper transaction after that. At this point, we're waiting pretty simply. We're just waiting for an outcome of that litigation that will happen over the summer. In terms of what happens next, it frankly depends on what happens over the summer, to be honest with you, Wamsi. I think we announced yesterday, we had our earnings call yesterday. We had a revenue and EPS beat in the quarter.
We narrowed both the revenue range and also the EPS range for the year. We also announced that we're going to do our security analyst meeting in early October. Obviously, once we come out after the deal announcement, we would go straight into the security analyst meeting where we would give you all that insight.
Okay, okay, great. If we kind of think about the broader demand environment, I mean, you just noted you reported earnings yesterday. How are you seeing the broader macro environment? Maybe you could just layer it in terms of all the concerns and the uncertainty that this tariff environment has created. What are you seeing as you talk to customers? What are you seeing in the demand pipeline? Maybe any things to highlight within the business?
Sure, absolutely. I think, as we said yesterday in our call, the demand environment remains relatively in line with normal expectations. Our linearity for the quarter was fairly typical, what I'd describe with normal seasonality. I would say that the start of the quarter demand was somewhat uneven, really just due to the fact that we were in the midst of, I think, a very dynamic situation with tariffs. As you would imagine, that created some level of instability as folks were trying to adjust to really what is the tariff environment. Now, obviously, as the quarter progressed, the tariff environment and the whole debate on reciprocal tariffs became much more clear. I think that's when we saw things start to settle into a normal rhythm. At this point in time, around pipelines, I'd say our pipelines are solid.
There's nothing particularly unusual across any of our businesses. The networking business, in fact, is looking like it's at a point of good health. We've seen strong pipeline performance there. Similarly, across our AI business, I think we gave some color and context around the pipeline is, in fact, multiples of our backlog. That's how we're sort of seeing the business right now, Wamsi.
Okay, okay, that's helpful. Maybe just pivoting a little bit to your earnings and the guidance that you spoke about yesterday, you narrowed the range a little bit. I think revenue came down about a point in constant currency. Earnings went up $0.08 at the bottom. Can you just talk about some of the puts and takes there? The 1% decline in constant currency was the primary driver of that.
Yeah, sure. Let me walk you through revenue, then I'll walk you through EPS.
Okay, great.
On the revenue side, you could imagine we're halfway through the year now. We've got much better visibility to the pipeline and how those deals are going to sort of move through the back half of the year. The narrowing of the range on the revenue was really driven by one simple factor. It was really just those AI deals. As you know, those AI deals are very lumpy. They're non-linear. They rely on customer acceptance of data center readiness. We were actually fortunate in Q2 that we had one customer that was more ready. We saw that deal pull into Q2, and we actually had 10% more AI revenue in Q2. What we've just seen is that now, as we get closer visibility to the back half of the year, some of those AI deals, those customers are not quite ready.
It's really just a reflection of customer readiness on AI or revenue. On EPS, obviously, we had a beat in Q2. We passed some of that beat through, which is what you saw in terms of lifting the midpoint up to $1.84. Part of that was driven by tariffs. The tariff exposure originally was $0.07 for the year. In the quarter we just had, we expected $0.04. It was only $0.02. We've been able to sort of work through that with some compliance opportunities on tariffs. We passed through that. Obviously, as we're halfway through the year, we narrowed the range on EPS because we have just more line of sight to the back half of the year.
On that AI commentary around push-outs, maybe just a step back on AI for a second, right? We've seen some of your competitors talk about very big sort of backlog and revenue numbers. HPE seems to be very deliberate in where it's playing. Maybe it'll be helpful to contextualize and say, what are these opportunities that you're targeting? What should people consider your addressable market? Maybe what are some of the margin, whether it's rates or dollars, that you feel like are metrics that are acceptable for you in your business to achieve?
Let me just walk you through how we view the market and then sort of give you some color around how we think about the business in terms of profitability. The AI market, as you know, is primarily driven by model builders. We sort of look at it in four key segments. First of all, model builders. Secondly, CSPs. Third, sovereigns. And sort of fourth, enterprise. I'd say on those model builder deals, they're very large deals. They're like big whales. They will absolutely distort your pipeline and distort your revenue recognition. Because when they happen, it'll obviously be a tremendous impact to revenue. Those deals come in into our desk, and we obviously look at them very carefully. We have a framework that we use around large deals, and we look at them and scrub them and make the decisions around those deals.
This also has a big impact on working capital. We have to take all of that into consideration. As you look at CSPs, fairly similar sort of nature. I'd say the two that we have perhaps a right to win and a right to play that we see potentially being better profit pools are sovereign and enterprise. Sovereign is really a reflection more of the supercomputing heritage we've had as a company. We've had long relationships with a lot of governments over the course of the years. It gives us naturally an entree into a lot of interesting sovereign opportunities. Then enterprise. In fact, I think Antonio did comment yesterday that we looked at the pipeline, and the pipeline definitely had a much better mix of enterprise this quarter than what we've seen historically in the past.
We're seeing some level of increasing maturity in terms of adoption in the enterprise. If I step back and sort of just look at this AI market, what I would say is a year on, if I look back when we had this discussion a year ago, we probably didn't see as much diversity in the pipeline geographically. Today, there's definitely a lot more international nature, sort of sovereign type deals that we just didn't see. Secondly, there is a maturity level that's starting to happen in the enterprise. Still a long ways to go yet, but definitely enterprises are adopting not just GenAI, but agentic AI, for example, and looking at how to drive both their business models and business efficiency through AI. That's how we see the market.
As I said, I think we are playing in model builders and CSPs where it makes economic sense. We will absolutely have a right to win in enterprise and sovereign.
Which of those two, which one do you think comes to fruition faster? We've heard of some large sovereign deals in the Middle East, for example, being spoken about recently. Would you say that as you think about those deals coming through, is that traditionally, first of all, where your supercomputing has played historically? Are those the customers where you expect to have the initial traction? Are you doing something to build out other areas to tackle within sovereign?
I’d say it’s a combination of both relationships, which obviously we’ve had for decades, frankly. Secondly, technology. We’ve had a long history. I think we spoke about this at AI Day last year around direct liquid cooling that came from that capability. It’s a combination of both, frankly. I mean, a lot of these governments around the world have worked with us for many, many years. We have a natural foot in the door, which is helpful. You’ve got established credibility. Plus, they’ve just seen the historical performance of supercomputing. We run some of the fastest computers on the planet. You’ve got proof points in terms of technical capability, know-how, and then the technology that, frankly, is required to run some of the newer NVIDIA GPUs.
Okay. And then just in terms of relative pace of adoption between sovereign and enterprise, would you say one faster than the other? Any color you can share on that?
Yeah, I'd say, look, there's very different drivers. The reason for sovereign sort of adoption is typically related to countries and states and around what they want to do and how they want to develop AI in those countries. In some cases, they're developing data centers to actually just promote AI within that country. In really being able to build a network and allow startups, et cetera, to really start to grow. Enterprises, we're seeing enterprises look at AI through a lens of not just productivity, but business transformation. I think that is very a compare and contrast to a year ago. We weren't even discussing agentic AI a year ago. It was probably something that was out there in the future. Today, it's real. We're implementing it actually in my own team. I know we're going to talk about that later on here.
It's become a lot more real and current than it was a year ago. Different drivers for companies and governments, but they're both starting to pick up some pace. It's going to take time.
Okay. Maybe it's a good segue to talk about just sort of what your initiatives are within HPE and how you're using AI. I think you made some comments yesterday on the call as well, but would love to understand what is how did that decision process evolve? What did you look at to say, like, yes, we need to invest X amount of dollars. This is the ROI we're going to get. Any color you can share on that?
Sure, absolutely. In fact, I'm just pleased to let you all know, as a CFO, I leaned in on an opportunity to drive with Deloitte a what we call Zora AI, which is a C-suite AI platform specifically designed for CFOs. We're actually collaborating to put it on our own architecture, actually on PCAI. We see it as a unique opportunity to really drink our own champagne and be able to talk to customers about not only how we're using our architecture, but how we're deploying AI inside of finance to actually drive better reporting, greater productivity, and frankly, better accuracy in reporting. In the midst of that, super excited about what that would bring to my own team because I just see that the ways of working can be substantially improved and increased. It's a great opportunity then for us to embrace inside the company.
Particularly what I'm excited about is actually I go to sell that externally to other CFOs. I've got to tell you, I've got a lot of calls. I'm not sure I'm quite ready for sales yet, but there's just a lot of interest from other CFOs who are trying to figure out how do I democratize AI in my organization? How do I keep it safe and secure? That's where our PCAI platform makes it easy for CFOs to get started. That's just one example. Honestly, inside the company, we're looking at opportunities, everything from customer service. We have a very large service organization, huge opportunity there that we are starting to embrace around using AI and helping us really take a lot of the mundane work out of sort of service organizations.
Frankly, even my earnings call yesterday, Wamsi, we use AI to help us prepare the scripts. Hopefully you saw some good scripts there yesterday. We actually even use AI to help us get better and more precise on our earnings calls. I actually predict your questions.
Oh, that's great.
That's quite not there, Wamsi.
Yeah, yeah. I wasn't on the call yesterday, unfortunately. I always get stuff here, but I'm sure you'll try to predict it for the next one.
Exactly.
Maybe a Juniper question, right? You mentioned about cash conversion and sort of the intensity of doing business around AI sometimes can have a materially different cash conversion cycle. Can you just flush that out a little bit? Just maybe compare to an industry standard server business that you're pretty large in. How does that cash cycle differ for AI from ISS?
Yeah, look, I think I'll start out by saying there's no doubt that the AI business is more working capital intensive. I think you've seen it from all the players in the industry. It has a whole different dimension because of the structure of the BOM, frankly. Also just the size of these transactions. They're very large deals. If you are working with model builders, those deals can be significant in size and scale. All of that ties into working capital. I'm going to sort of distinguish that from our leasing business, which is a whole different sort of operating structure. It's absolutely one of the considerations that we look at as we look at these large deals and large transactions, what impact it has on working capital.
I think as we commented on cash flow yesterday, I reiterated we're approximately $1 billion for the year. That's really attributed just to the dynamics around working capital.
Yeah. Maybe on that, right? The first half of the year, you used $1.7 billion in free cash. You're guiding to a billion. You got another $2.7 billion to deliver here in the back half of the year. Some of it is related to this large AI transaction that you noted. What are some of the other puts and takes? How do we get to $2.7 billion in the back half?
I think you probably are familiar that historically, HP, HPQ, and E actually have a very seasonally back half loaded in terms of free cash flow. The company generates most of its cash flow in the back half of the year, which is into Q3 and Q4. We typically see that occur from here on. As you mentioned, we'll have that large AI transaction that'll move through inventory into revenue recognition in Q3 in this current quarter. Obviously, we just see the sort of drivers, the fundamental drivers of all of our business that drive the volume in the back half of the year.
Okay, okay.
All of that contributes to cash flow. I think I noted in the prior call, obviously, we've got some restructuring around the cost program that we expanded yesterday to include catalysts. Some of that is also in our cash flow this quarter as well.
Can we just talk about?
Some this year as well.
Can we just talk about restructuring for a second, right? I mean, if we just step back for a second, I know over the years and even prior to your being at HPE, there have been plenty of restructuring programs, right? When you look at the overall sort of flow through of those, maybe to the bottom line, externally, it's not entirely obvious how much is flowing through. Can you give us some context as you think about either this expansion of this most recent plan? How much should investors expect flows through to the bottom line versus things that it might be offsetting because of external headwinds?
Yeah, look, I mean, I think what's different here is that we're very focused on the impact of the savings. I think I mentioned on the call yesterday that actually our headcount hit 59,000, which is the lowest it's been since an independent company. To your point, you need to see those impacts flow through in terms of OpEx structure and then obviously through to profitability. [Monolocally] focused on that is what I would say, Wamsi. We did announce a broader catalyst plan that includes beyond not just workforce transformation, but other structural cost initiatives aimed at efficiency, portfolio, and even using AI more effectively for productivity. We'll give a broader sort of update on what that does to our longer-term outlook when we get to the security analyst meeting in October.
Okay, okay. Yeah, looking forward to that. Maybe to step back and just talk about tariffs for a second, right? It has created a lot of volatility. I mean, it has been hard for people to kind of really understand where the bogeys are, where to move things even around. Just talk to us a little bit about how you are handling this uncertainty from an HPE perspective. Where are your assets? What can you do strategically if things get more fluid from here?
I think I'll underscore your comments about uncertainty. It's been, for all of us, I think, an unprecedented time in terms of that. I would say that when we guided last quarter, we were probably one of the first companies out the gate that had to incorporate tariff guidance. At the time, we hadn't had the reciprocal tariff. We estimated about $0.07. Post that original announcement, we availed the opportunity of the USMCA compliance, which really initially was our server business, but we were able to take advantage of that in terms of a lot of the activity in our server portfolio. We've mitigated some of the tariff exposure. As I commented, it's down to $0.04 now for the year.
I think the key, though, if I step back and look at this in the more broader context, is having a globally distributed supply chain. We were fortunate that we do have that. We have a supply chain that's well distributed and not highly concentrated in one particular part of the world. I think many of us learned through COVID that you need to have resiliency and you need to have a much more sort of distributed operation given what many of us faced during that period of time. Obviously, you need to have flexibility in that supply chain as well to sort of deal with some of the dynamic nature of what we've experienced most recently. We've been able to navigate that, I think, fairly well to the point that we're able to reduce the tariff exposure.
I was pleased to be able to update that and pass that through in the guide yesterday.
Just from a physical location of where those assets are and how you're sourcing, how much of your base is USMCA compliant versus not? Is there a move to, are you doing incremental moves within your supply chain to any particular countries or regions that you think is notable?
Look, I'd say we're always looking for opportunity, Wamsi, to sort of expand our global footprint. Certainly, the USMCA compliance was important for us in terms of being able to mitigate some of the tariff exposure. We're constantly scanning the market to look at ways to optimize our supply chain.
Okay. Marie, maybe just thinking about industry standard servers, where are we in our replacement cycle with respect to that? I mean, I think at the start of this year, most of us anyway felt that we were looking at a very strong IT spending year after some pent-up demand and sweating off assets for a few years. It's improving, not just maybe quite to the degree that we had hoped it would. Any color you can share on how you see that evolving from here?
Yeah, no, I think, first of all, we announced our Gen 12 transition. I would just sort of comment that that's been going very well. We're seeing higher AUPs. I think you're absolutely spot on. We sweated assets a lot longer, particularly, I think COVID was one of the contributors to sweating those assets during the period of time. I know folks had a lot of Gen 8s out there. The newer Gen 11, Gen 12 provides an opportunity for folks to really modernize their data centers. We've seen that certainly play out. I would say from a traditional server perspective, albeit, this last quarter, we did see some unit volume adjustment as we were moderating pricing. We still do expect to see unit growth in the back half of the year.
I think that just echoes your comments around the fact that we still have good expectations around the growth and strength of data center modernization.
Okay. Would you say that just from an enterprise customer standpoint, I think some of your peers noted a slightly weaker close to the quarter, a little bit of uncertainty, whether some noting either public sector weakness, some noting some European weakness, North American weakness.
Every weakness everywhere.
Yeah, weakness, weakness across the board. What did you guys see as you closed the quarter out? I mean, I think you actually said you saw some uncertainty at the beginning of the quarter.
Correct.
How was linearity towards the end of the quarter? How has it been early days into this quarter?
Yeah, look, first of all, say we're all dealing with macro uncertainty. I think everybody, and I think there's one call that folks haven't used those words. I think I contrast the beginning and the end of the quarter. When we started the quarter, we saw more uneven demand. When we ended the quarter, it was probably more consistent with normal expectations. We went back and looked at linearity. There was nothing out of context in terms of what we've seen in terms of normal seasonality. I'd say ended the quarter with normal sort of seasonal patterns and a relatively strong pipeline in terms of orders.
Okay, okay. As you're looking at your own internal spend metrics, given the uncertainty more broadly in the macro, are you changing anything internally as you're planning? As most companies plan for their fiscal year budgets, I mean, I think given all the recent changes, everyone's reevaluating to some degree. I'm just curious to hear how you guys are thinking about it from an HPE perspective.
I can tell you I'm always looking at it, Wamsi. Probably too much. But I'm always looking at ways to drive cost structure in the company. In fact, last quarter, you recall we actually announced that we were going to undertake a significant workforce transformation, exiting around 5% of our workforce. I think to a certain extent, we got ahead of it. We saw the tariffs coming. We did not want to wait. We got ahead of, I think, a lot of that uncertainty. Frankly, we're just continuing on from there. That is why we announced Catalyst, which includes a broader suite of efficiency opportunities, both at driving cost structure, improving COGS, and frankly, driving top line. Nothing's off limits here is what I would say.
Yeah. At the beginning of the year, in Q1, there was some mis-execution around pricing and inventory and a few other things that happened in the quarter. Can you help me to maybe give confidence to investors that that thing's kind of in the rearview? What are some of the actions that you've taken that won't cause issues like that in the future? Maybe just to share, what were some of the things that maybe failed from a process standpoint that you could fix?
Sure. If you look back at what caused those execution issues, there's really just two big buckets. One in traditional compute, which is around pricing and discounting management. The other was really in the AI space where we were managing through the transition in the GPU cycle in terms of AI inventory, which I think is an industry-wide acknowledgment. I think we worked through a lot of that through the quarter. We've still got more work to do. The entire industry is processing that GPU transition. On the traditional server side of the house, what happened there was really a combination of pricing and discounting controls. We have put in place much more stringent controls around our bid desk during the quarter. We did have backlog that was priced at a different point.
Henceforth, that's why we guided the way we did into Q2. The expectation is that Q2 would be the trough of server margin. I think you saw that happen. We've guided to sequential improvements, Q2 on Q3, Q3 on Q4, exiting the year with server margins approaching back to the sort of bottom of the range that we had guided of 10%. You can see that evolution of margin as we've improved the controls around pricing and discounting. I can just say we've got very, very stringent controls there today.
Right. On that trajectory to 10%, you also noted in Q3, you're going to have this very large AI deal. We should really expect that to be more Q4 loaded in terms of magnitude of quarter-on-quarter improvements as we think about the next couple of quarters.
Absolutely, yes. As we said on the call yesterday, we do expect a large AI deal to ship in Q3. That was actually driving both the revenue and the operating margin profile for Q3, as you correctly said.
Right. In terms of attach of the portfolio to your AI opportunities, how do you think about the broader attach of services? Obviously, much higher margin, much more relevant in enterprise and sovereign than it is at tier two CSPs. In aggregate margin terms, is there a number you would be comfortable sharing on what could be the operating margin when you think about some of these large server deals?
I think we're not going to get into the specifics on those deals and the operating margins. In terms of services, what I would say, actually, we do disclose now in our investor deck the mix between product and services. If you look actually at the most recent disclosure, you'll see that the mix of services has grown quite nicely from a sequential perspective. It is trending in the right direction. We're starting to see services mix grow. That is what gives me sort of pause for the opportunity, particularly in sovereign and enterprises, where those services are really going to perhaps be a greater asset to those types of deals. Model builders may have a lot more of that capability in-house, whereas sovereigns and enterprises who do not necessarily have all of that capability will rely more heavily on our services capability and portfolio.
Maybe to just step back and think about sort of the from a valuation standpoint, right? I mean, it just feels like you guys have a portfolio that, I mean, this last past fiscal year generated $2.4 billion in free cash flow, went to $2.4 billion free cash flow. Obviously, you've had this year with more volatility and things that were unexpected. If the portfolio independently is able to generate something to the order of that, and you look at your financing debt, which is largely backed by very high-quality receivables, you really are trading at very low multiples. Now you have a company that actually just delivered six, seven points of growth. You're projecting something higher than that for the year.
As we think about putting all those pieces together, it feels that investors are not giving you credit for what you can deliver here in the next several years. I know you've been super focused on costs, and you've obviously shown and delivered that in your prior roles. When we put all those things together, why is this not a great time to kind of go out there to do significant amounts of buybacks? Because I think from a value creation standpoint, that does come up as a fairly significant option.
What I would say in terms of the portfolio, you're absolutely spot on. I mean, there are drivers in each of these businesses that I think are very positive. I think we've seen the momentum in hybrid cloud. We've seen the momentum in service. Discussion we've had on AI. I think the great thing about networking, we're starting to see that there's definitely a shift in the whole industry. What's interesting is that each part of the portfolio is very well positioned in terms of drivers. Yes, I would agree with you. The stock is certainly undervalued. What I would say, I think we started the conversation here, and I think it looks like we're going to end it there, which is we have a very important transaction that's coming up over the summer, which is the Juniper litigation outcome.
We're waiting for that as a sort of key next step in terms of how that would impact our capital allocation strategy going forward. I'd say stay tuned for Sam, Wamsi.
Yeah, okay, fair enough. Marie, we're just about out of time. 30 minutes is just too less.
I know, it went quickly.
Pick your brains on all these important topics. Thank you so much for being here.
No, thank you. Thank you for the opportunity. It was a pleasure. Thank you, everybody.