HPE Boosts Networking Forecast on AI Demand

0:02Speaker 1:Bloomberg Audio Studios. Podcasts. Radio.

0:06Speaker 2:News.

0:08Speaker 3:You're listening to Bloomberg Business Week with Carol Masser and Tim Stenevek on Bloomberg Radio.

0:16Speaker 4:We're going to keep our conversation going here about AI, the build-out that's being caused as a result of it. as we talk to the CFO of Hewlett Packard Enterprise, HPE.

0:24Speaker 3:She joins us now.

0:25Speaker 4:That's Marie Myers. Great to have you with us. I know that you were listening to some of that conversation there. So maybe I could begin just by getting your perspective on sort of where things stand. Of course, you have heard those who are anxious about what this might mean for humanity, those on kind of the doomer end of the spectrum here. But you are integrally involved here in the wider build out of the systems that are buffeting and powering this AI renaissance. How are you feeling about this moment? How should we think about the moment from your vantage?

0:51Speaker 1:Well, good afternoon, Dave, Christina.

0:53Speaker 5:It's an absolute pleasure to be with you this afternoon. I've actually just come from our investor day here in Sunnyvale on our networking business. And clearly, you know, we're in a moment of time where there's just phenomenal momentum driven by AI and architecture and also just in terms of how AI is really transforming work and our lives. I mean, I use AI personally every day and I've been a very strong advocate in my role as a CFO of HPE and actually building agents to help us run our own business. And we've seen tremendous success and transforming workflows, frankly, that we've had for decades that perhaps were somewhat useful, but today I'm seeing just enormous change and it's really benefiting the way we work. However, I'd say what's really important is the approach around responsibility. And as a company, we've had an all-star.

1:45Speaker 1:I would say, about responsible AI.

1:49Speaker 2:I do want to ask you, I mean, I know you've made some big investments and some collaborations are upcoming and we're going to talk about that. But when you look at the effectiveness of AI in your business, you said you use it every day. Do you draw a distinction between the usefulness of generative AIs and, you know, agentic task-based systems?

2:08Speaker 1:Absolutely.

2:09Speaker 5:And I would actually even add into that. the lawsuit you've given is also automation. So myself and my team, we typically look across the board spectrum of use cases, and then you have to sort of figure out, you know, which one is the right one to use for the different workflows. And in terms of the benefits, you know, I would say clearly the agentic capabilities have really sort of taken off in the last 12 to 18 months. I think it's a moment many of us were waiting for as enterprises are perhaps laid, you know, the early stage foundations with generative AI. I almost would start with automation, generative AI all the way through now where, you know, agentic workflows and the development of agents is becoming much more common in the enterprise. And clearly, I mean, at home, honestly, I have my own agents that help me out every day just to sort of keep keep my own life managed.

2:57Speaker 5:But at work, I rely on those agents for some really critical, as I mentioned, really critical workflows. We redesigned a call that we had on a Monday. It used to be a 300 plus PowerPoint deck that would take my team weekends to do.

3:11Speaker 2:I was going to say, listen, I'm an AI skeptic, but if you can solve that, I'm right there with you.

3:17Speaker 1:Exactly. Exactly. And it takes a lot of time away.

3:20Speaker 5:And actually what I think is really more important is that It made the conversation so much easier with the team.

3:27Speaker 4:Marie, you mentioned your investor day, and I know there was a big deal announced here, a $ 1. 2 billion order from a cloud computing company. Talk a bit about that, the contours of that deal, and indeed what the message of that deal is sending as you see it.

3:41Speaker 5:Yes, look, that deal is really important for us because it is our first big AMD sort of Clearly, for us, it's the entire AI system, which includes both our networking switch all the way through to software and other hardware. But what's really sort of probably taking everybody by note today is the the integration of the Helios switch into that rack. And actually at the Investor Day, one of my favorite parts of the day was actually, you know, folks seeing this switch. It is literally, if you see it, like a work of art. It's incredibly wide because this rack is much wider than normal racks. And there's 1,700 copper cables that are interwoven in this switch. So I never thought I'd say it like that, but the general manager of the team here has done an amazing job to produce... what I think is going to be a work of art in switch history. But it also just underscores the performance needs and.

4:39Speaker 1:Sort of the levels that are really rising.

4:41Speaker 5:In terms of performance just for AI itself, that you need to actually have a switch that works like this and runs like this. And I'm so excited that, you know, with the acquisition of Juniper, we were able to develop this technology using the best of what HPE had prior to the acquisition and then taking the Juniper engineering and coming to market with A & D to, announce this deal.

5:03Speaker 2:Talk to us about how this Helios AI rack, and I definitely totally understood what that did and.

5:09Speaker 3:All those lovely words you just used.

5:10Speaker 2:But as you look at where you want to take this technology and where you're pushing the business, how does this factor in?

5:15Speaker 3:What does it give you? What does it do?

5:17Speaker 2:How does it fit into your business plan and how you're integrating AI?

5:22Speaker 5:Yeah, look, the real sort of power of this moment is the fact that as a company today, HPE now has a firm entree into the data center. Before the acquisition of Juniper, I would say that we were very well known in the enterprise for our networking business, but we didn't necessarily have the same sort of entree into the data center. A combination of this deal here now with Vulture, who we've had a relationship with several years, and another deal that we announced back in our earnings call in our last quarter with Oracle, now reposition HPE networking right in the heart of the data center. And that's what I'm super excited about because we have a chance to really participate in AI infrastructure build-outs in a way that we didn't have prior to the acquisition of Juniper.

6:12Speaker 4:I wanted to ask you just who you're taking market share from and sort of where you see opportunities in the wider market going forward here.

6:17Speaker 3:So is it Cisco? Is it NVIDIA? Sort of situate yourself in that ecosystem, if you would.

6:24Speaker 5:Look, it's a very broad and competitive ecosystem. And today, actually, what we did say at the investor day, another one of my sort of favorite highlights of the day was not only did we raise our revenue outlook, we raised our cost synergies, We also raised our AI orders for the quarter.

6:40Speaker 1:But we underscored all of.

6:41Speaker 5:That by saying we plan to take market share across the different lines of business and networking. So a really powerful moment for us leaning in. And, you know, our expectation is that we're going to continue to drive share gains across the portfolio.

6:57Speaker 2:I mean, one of the things we talk about constantly in tandem with the AI discussion is, of course, the energy factor. How are you looking at that as you get more invested into this? I was reading up, there are some things you're doing to try to make these components, some of these systems, a little bit leaner in terms of their energy consumption. Can you talk us through some.

7:14Speaker 1:Of those efforts? Yes, and I'd say that.

7:18Speaker 5:As a company, we've had a long history in a particular technology called liquid cooling, which is really designed to help to cool down the key elements in a rack. Liquid cooling was traditionally used a lot in our server and supercomputer business.

7:33Speaker 1:Our heritage goes all the way back.

7:35Speaker 5:We actually power six out of ten of the world's largest supercomputers. So we had to learn how to cool those computers over the time. We were able to take that technology now and actually apply it to our networking switches. And so that's a massive leap forward. In fact, we were one of the first companies to come to market with a liquid cooled switch. And so sustainability has always been at the heart of HP. I've been with the company for nearly 30 years. And I think we have, as a company, always had it as a North Star and really been focused on how we could just not only lead in the way we do things, but actually take it all the way through to our engineering and our products. And these are great examples, living examples of what we do day in, day out.

8:18Speaker 2:When you talk about sustainability, especially when you talk to politicians or shareholders, in the last couple of years, are you changing how you're framing that? I know we were talking to the CEO of Trane who was saying they're not doing it just for the altruism of it, which I think was the conversation a couple of years ago. He's like, it's better for the business. It's more efficient. It's better for our profit margins. Is that how you address that? Not just from a it's better for the planet kind of perspective?

8:45Speaker 5:Look, I'd say we've always, like I said a moment ago, we've always had a history. I would say it's been in our DNA, but it's the right thing to do. And, you know, what we typically try to do is guide our customers and help them navigate that journey around power efficiency and as they think about standing up data centers and look at how they actually might want to redeploy and rebuild those data centers. So it's an important attribute that we absolutely focus on and navigating our customers is clearly an important part of what we do.

9:13Speaker 3:Maria, a couple of questions about servers, if I may. Really?

9:16Speaker 4:Clear to me that you are an AI networking company now. Why not be more aggressive on the AI server front? That's my first question. The second is, Obviously, the traditional business for HP has been traditional service. That's been performing well as well. What guidance can you give us? What sense can you give us about the trajectory of that part of the business going forward?

9:35Speaker 5:Yeah, so maybe I'll start with the traditional server part of the business. So last quarter, we commented on the incredible performance that we've seen in our traditional server business. In fact, the demand has been so strong. Our orders in the last earnings were up 70% on a year-on-year basis. And what we're starting to see, which I'm super excited about, is the move into inferencing. And it's really starting to be a tailwind for our traditional server business. In fact, I spoke about a very large deal. It was about $ 3. 5 billion, where that enterprise customer was really focused on using traditional compute for inference. And I would say we see it in the way we run our own company.

10:12Speaker 5:In fact, we made a a conscious decision to go on-prem with our own AI architecture. So, you know, clearly inference is going to become an increasingly important tailwind in our traditional compute business. And as we pivot and think about our AI business, we have been focused on the part of the portfolio that we felt that we could play and win. And that was really the enterprise. We've been in the heart of the soul of the enterprise for decades. And then also, as I mentioned earlier, we power six out of 10 of the world's largest supercomputers. So, Sovereign has been another place where we have played very successfully.

10:45Speaker 5:So our AI framework, particularly for large servers and AI servers, has been really guided by, you know, how can we help those enterprises scale and get ready for this AI adoption cycle? And then similarly, you know, how can we, you know, leverage the long history and relationships that we've had in Super Compute into the Sovereign space? So enormous tailwinds in both spaces. And clearly for us, it is never been a better time to be.

11:10Speaker 1:In this business.

11:12Speaker 2:You've boosted some of your guidance. You have seen really solid growth, but I'm wondering if you think this is sustainable. Can you talk about the durability of this kind of growth through 2029, given all the factors we've been talking about all day, the uncertainty of AI going forward in some markets, the pullback in some of these IPOs? Are you confident that this is a trajectory you can sustain for years to come?

11:34Speaker 5:I think you have to look, to really get to the point of sustainability, you need to look at the drivers of what's going on in the data center.

11:40Speaker 1:And I'd say there's sort of two powerful drivers.

11:43Speaker 5:One, there clearly is a lot of sort of compression on data centers themselves. Companies, organizations are having to modernize, improve the footprint of the data center, improve performance. So in itself, companies are looking for the latest technology so that they can improve the performance and the outputs of those data centers. Secondly, you know, I'd say we're in the early innings of AI adoption. I think enterprises have just started to move from what were last year, all the conversation was about proof of concepts. Today, you know, I speak to a lot of CFOs, CEOs and other AI leaders.

12:19Speaker 1:In the market.

12:20Speaker 5:Today, people are actually using agents in the enterprise. We're using digital agents in our personal life. You know, that sort of pivot has.

12:28Speaker 1:Only just begun.

12:29Speaker 5:So I would say we're in the early innings of AI adoption.

12:33Speaker 1:So we see.

12:33Speaker 5:Those tailwinds as providing durability over the long term for our company.

12:38Speaker 2:All right, Marie Myers, she is HPE's CFO and joining us today live over soon.

12:43Speaker 1:But thank you so much for being on.

12:44Speaker 3:Great to see you.

12:44Speaker 1:We really appreciate it.

12:45Speaker 3:On Investor Day. Thank you very much.