Key insights
- Hewlett Packard Enterprise (HPE) reported a significant beat on both earnings and revenue, driven by exceptionally strong demand for its AI servers. The company's stock surged as it raised its 2026 outlook and provided 2027 guidance, signaling robust growth prospects in AI infrastructure. This performance highlights the broad market's focus on AI adoption and the potential for companies enabling this trend to outperform.
Hewlett Packard Enterprise (HPE) is holding onto major gains ahead of Tuesday's close after topping fiscal second quarter estimates for adjusted earnings ($0.79 per share vs. forecasts of $0.54) and revenue ($10.68 billion vs. forecasts of $9.74 billion).
The tech company's stock moves are driven by the growth of its AI server business, now being considered a new contender in the AI infrastructure race.
HPE CEO Antonio Neri sits down with Yahoo Finance Executive Editor Brian Sozzi to talk more about the earnings results, 2027 guidance, AI infrastructure, and supply constraints.
This was HPE's biggest earnings beat seen since 2018, on top of record revenue.
HPE shares are surging after blowout quarterly results. Soaring demand for AI is propelling growth for that company. Yahoo Finances executive editor Brian Sozzi has been speaking to the company's CEO Antonio Neri on the latest quarter. Take a listen.
What happened is that we delivered an exceptional quarter, Brian. And that quarter was marked by record breaking results across the company. We posted obviously, uh, record breaking revenue. We significantly exceeded our non-GAAP EPS, and we delivered a very strong free cash flow outlook, but most importantly, we grew orders much faster than revenues pointing to the very strong demand for our portfolio, which I think is perfectly in sync with the market that we see today across networking cloud and AI. And because of that strong performance, the record breaking backlog we have and the visibility into the demand which is durable, we are actually not only raising 2026, um, you know, the outlook which actually is pulling by two years, the 2028 commitment. So think about that. and then also guiding 2027. So very, very proud of this moment in time and is the outcome of a very curated strategy that's paying off and the hard work of 65,000 employees. So, very excited about today.
Antonio, did companies suddenly wake up in this quarter and realize that they were behind in building out their AI infrastructure?
No, what I say is that, you know, when I speak to customers, which is more than 50% of my time and last week I was in Chicago hosting round tables with customers and partners, channel partners are very important to us. You see that nobody wants to be left behind in the AI adoption. A Gen AI has definitely accelerated that momentum because it's better for transforming processes than just adopting large monolithic models. I think it's a combination of mix of models between open and closed models. And the barrier to entry is much lower now. You think about what we're doing in the AI factory for enterprise, we make it very easy for customers to deploy the infrastructure on prime or in a colo under their control under governance with the simplicity of deployment and life cycle management. But also we bring the entire software ecosystem to them because we have this deep partnership with Nvidia where we actually integrated their software stack and our software stack to Green Lake. So clearly there is an acceleration, yes we always say Brian, you know, the future belongs to the fast and they understand this industrial revolution is powered by AI.
Why was it so important for you to put out fiscal 2027 guidance?
Because we want to make sure the the market understand this is durable demand. you know, and when I think about our record quarter backlog, the pipeline which remains multiples of our backlog, and the incredible moment that we have in networking cloud and AI, this is a a longer cycle. Despite the fact that supply continue to stay constrained. But in that guide Brian for 26 and 27, we factor the supply that we know we are getting in 26 and the capacity that we have locked in our long-term agreements with our suppliers as a part of 27. And we decide on a 90-day basis how to divvy that capacity based on the pipeline and the deals we're closing. So, this is a strong momentum for us in addition to the fact that the juniper integration is going well and the catalyst modernization using AI is going very, very well. We have 1,200 use cases in AI and 250 in production just for our company, which is helping our Opex and cost of sales constructure.
Antonio, do you envision is envision Is is AI infrastructure demand just likely to head up up and to the right over the next decade? I'm trying to figure out what what would that point look like? where is that cliff? because there was so much concern, I think coming out into this year that eventually we're going to hit a peak and it's going to fall off off a cliff somewhere and a lot of companies are going to get screwed. I mean, what do you how do you think about that?
I don't think at all about that. I mean, if you think about the number of gigawatts being announced so far, which is now over 150 gigawatts, give or take. And how much capital is required to build one gigawatt of data center and the number of GPUs in each uh, in each data center. If you make simple math, you know, it's between $60 and $70 billion per gigawatt between the building, the power, the cooling and the GPUs and any accelerator you put in or even CPUs, you know, as a ratio to GPUs. Uh, and you multiply that by 150. You're talking about trillions of dollars. And with the environment we live in, with where supply is constrained, we have a cycle here all the way to the end of the decade that will continue to accelerate. And as enterprise adopt these models uh in their workflows to improve productivity, I I think this we still so early in the cycle, Brian. So that's why, you know, you need networking to make GPUs productive. You need a lot of data, storage platform to manage the data. You need an immense amount of compute power.
Lastly Antonio, I was still uh a young buck entering college during the .com runup and then the bust. Um but I think you were, you know, out there managing, doing your thing uh in your career. Why is it why is this time different compared to that period? I get this question all the time. Everybody's asking me, Brian, is this a bubble? and the short answer is I don't know. So I'll just put the question to you.
No, I think we need to all realize that the market has changed forever. And this uh revolution that we call it, the industrial revolution is powered by AI. You know, you you said, you know, you were young during that time, I lived all the major technology transitions, the main frame transition, obviously, the the PC client era, the internet transition, the mobile transition, the cloud transition. And when you put in the perspective when the first telephone, think about it, the first telephone was available through the landline. It took almost a decade to get 50 million users connected. We got 50 million users using AI in one month. And so think about it, the scale and the acceleration we all live in. This technology will change society, obviously, will change the way we work. Um and now it brings other challenges, right? You know, making sure that it's equitable and everybody can get access to it. but reality, we have changed the market forever. and that's why looking into the past is not a reflection into the future. There are lessons learning in the equity market, I get that. but in term of build out of infrastructure, totally different.