CEO Antonio Neri: HPE Is Seeing A ‘Wave’ Of AI Acceleration
‘The HPC and AI business segments saw a wave of demand acceleration in the quarter as we converted on AI deal opportunities and shipped orders that leverage our unique end-to-end AI value proposition, from training to tuning to inference,’ said HPE CEO Antonio Neri. ‘As a result, we exited the quarter with the largest HPC and AI order book we have ever had.’
Hewlett Packard Enterprise is seeing a “wave” of AI and high performance compute (HPC) “demand acceleration,” helping to fuel a 122 percent increase in HPE GreenLake-as-a-service orders in the most recent quarter, said HPE CEO Antonio Neri.
“The HPC and AI business segments saw a wave of demand acceleration in the quarter as we converted on AI deal opportunities and shipped orders that leverage our unique end-to-end AI value proposition, from training to tuning to inference,” said Neri during a conference call with analysts Tuesday after Spring, Texas-based HPE posted better-than-expected quarterly results for its third fiscal quarter, ended July 31. “As a result, we exited the quarter with the largest HPC and AI order book we have ever had.”
HPE’s AI “momentum” also helped drive the company’s total order bookings to more than two times pre-pandemic levels, said Neri. He said the HPE AI pipeline continued to be “super strong” in the quarter. “Clearly the momentum in the business is significant,” he said.
HPE is even seeing AI opportunities with the large hyperscalar public cloud providers, said Neri, including a deal completed in the third quarter that has not begun to be built or shipped.
“The reason why [public cloud hyperscalers] are coming to us is because we have a unique amount of intellectual property,” he said. “Think about it as an open ecosystem with our networking fabric which obviously supports Nvidia—we have a fantastic relationship with Nvidia—but it also supports other type of accelerators and, depending on the AI workload, it can be a mix and match.”
One factor that could hinder the delivery of AI solutions is supply constraints, said Neri. “Supply obviously in the AI space is constrained,” he said. “No. 2 is the fact that when you deploy these deals you have to install it and drive acceptances, which means elongated times for revenue recognition.”
Even with the big AI wins driving GreenLake-as-a-service orders, HPE’s HPC and AI revenue during the quarter was $836 million, up one percent from the prior year, with 0.8 percent operating margin compared to 3.4 percent from the year ago quarter.
Going forward, Neri said he expects the AI momentum to accelerate, but with revenue recognition lagging the bookings. “You have to look at this as not just the next quarter but on a mid-to long-term basis, call it two, four and then eight quarters,” he said.
HPE Sees Big AI Wins In Education, Research
The AI gains are being driven by HPE’s complete lifecycle of solutions, said Neri. He singled out HPE’s unique AI software offerings and its Slingshot networking fabric, combined with what he called HPE’s “market-leading sustainable supercomputers” and its as-a-service offerings.
HPE’s open ecosystem of AI suppliers is also an advantage for customers who are turning to HPE for a “full spectrum of enterprise AI workloads and use cases, spanning large-scale model development, training, tuning and inferencing,” said Neri.
Among the big AI wins for HPE in the most recent quarter was a deal with the United Kingdom’s GW4 Alliance, which brings together the universities of Bath, Bristol, Cardiff and Exeter, to build a supercomputer that leverages the latest HPE Cray XC Supercomputers, HPE Slingshot interconnect and Nvidia GPUs, said Neri. “The system will provide researchers, engineers and data scientists purpose-built capabilities to train AI models and accelerate research in clean energy, drug discovery, medical diagnosis and astrophysics,” he said.
HPE was also selected by the Tokyo Institute of Technology Global Scientific Information and Computing Center to build its next-generation supercomputer, which includes AMD CPUs and Nvidia GPUs, said Neri. That supercomputer is aimed at accelerating “AI-driven scientific discovery” in medicine, material science and climate research.
In addition, Recursion, a biopharmaceutical company, is now using HPE’s AI software to scale its efforts, significantly speeding up its AI training across more than 25 petabytes of biological and chemical data, said Neri.
Although HPE saw a 13 percent drop in compute revenue to $2.6 billion, Neri said HPE anticipates “demand picking up in the coming quarters” as a result of customers seeking systems to run AI inference workloads. “Our new HPE ProLiant Gen 11 servers optimized for AI workloads are well-positioned for this growing customer need,” he said.
In the third quarter, HPE started shipping the Gen 11 servers, which boost HPE AI inference performance by more than five times over prior ProLiant models, said Neri.
HPE also last week expanded its enterprise tuning of AI inference solutions with VMware and Nvidia to accelerate generative AI deployments, said Neri.
GreenLake Partner Of The Year Sees Investments Paying Off
HPE has a “significant competitive advantage” in the AI market because of its long standing investments in AI supercomputers and software, said Pat O’Dell, managing partner for Clinton, N.J.-based CPP Associates, which was recently named HPE GreenLake North America Partner of the Year.
“HPE is ahead of the curve when it comes to AI and supercomputers,” he said. “One of the reasons CPP has invested so much in HPE is they have by far the most comprehensive product portfolio.”
O’Dell credited Neri, who was No. 3 on HPE’s Top 25 Most Influential Executives list, for paving the way for HPE’s big AI gains with investments over many years, including the blockbuster $1.4 billion HPE acquisition of Cray in 2019.
“Antonio is always moving to where the puck is going, not to where it is or has been,” he said. “There is no one that has a more comprehensive and as relevant a portfolio, whether you are talking about AI, as a service, private or public cloud and even traditional infrastructure like server, storage and networking. HPE is in the vanguard on all of them.”
CPP’s GreenLake sales are up triple digits with deal sizes increasing, said O’Dell. “GreenLake is becoming much more mainstream in our conversations with customers,” he said. “Seven or eight out of 10 customers are considering GreenLake. We’re definitely seeing more GreenLake momentum.”
Overall for the quarter, HPE reported non-GAAP diluted net earnings per share of 49 percent on a one percent increase in sales to $7 billion. The Zacks consensus estimate was non-GAAP diluted net earnings of 46 cents per share on sales of $6.98 billion.
Among the highlights of the quarter was a 50 percent increase in HPE Intelligent Edge revenue to $1.4 billion, the company’s fifth consecutive record quarter for intelligent edge sales.
HPE’s Intelligent Edge business with Aruba networking now accounts for the largest segment of HPE’s operating profit, accounting for 49 percent of total operating profit.
HPE GreenLake business, meanwhile, accounted for a $1.3 billion in annualized revenue run rate in the quarter, a 48 percent increase from the year ago period. HPE’s GreenLake total contract value is now nearly $12 billion.
HPE said in the wake of the strong results it is raising its full year non-GAAP diluted net earnings per share guidance to be in the range of $2.11 to $2.15, up from $2.06 to $2.14.
HPE shares closed Tuesday up five cents at $16.84, but were down in after hours trading by 19 cents at $16.65.
“While the broader IT market is still pressured, demand for our products and services grew sequentially in the third quarter across all key segments of our business, driven by high growth areas like AI and HPE GreenLake,” said Neri. “We continue to see strong interest in our AI and supercomputing offerings from enterprise customers who are incorporating artifical intelligence into their businesses. This is translating into significantly higher demand for our HPC and AI business segment as customers discover HPE’s unique capabilities to power unprecedented levels of performance for AI at scale.”