HPE ‘Significantly’ Expands Nvidia Partnership To Help Businesses Adopt GenAI Faster

The move is part of Hewlett Packard Enterprise’s bigger push to develop what it calls “AI-native” infrastructure, software and services offerings that meet the unique and computationally intense requirements of a wide range of AI workloads.


Hewlett Packard Enterprise said it’s doubling down on its partnership with Nvidia to introduce an “enterprise-class, full-stack” solution meant to help businesses adopt generative AI applications faster.

The hybrid cloud powerhouse announced the “significantly” expanded partnership with Nvidia—the dominant AI chip designer whose revenue tripled last quarter due to a boom in GenAI spending—at the HPE Discover Barcelona 2023 event on Thursday.

[Related: Nvidia’s H200 GPU To One-Up H100 With 141GB Of HBM3e As Memory Race Heats Up]

“It’s very, very clear to us that the combination of Nvidia and HPE in this enterprise computing solution for generative AI represents the fastest path for enterprises to embrace and transition to deploying generative AI,” said Neil MacDonald, executive vice president and general manager of the HPE Compute business, in a briefing with journalists and analysts.

The move is part of HPE’s bigger push, outlined at the event, to develop what it calls “AI-native” infrastructure, software and services offerings that meet the unique and computationally intense requirements of a wide range of AI workloads, from computer vision to generative AI.

“We think the next decade is going to require full-stack thinking from hardware all the way up through software layers as organizations lean into the deployment of AI applications,” said Evan Sparks, vice president and general manager of AI solutions and supercomputing cloud at HPE.

MacDonald said the new AI solutions present a “tremendous opportunity” for HPE’s channel partners to help businesses figure out how to take advantage of GenAI.

“Channel partners are essential to everything we do. We grow with and through our partners in serving our customers in the market, and the AI opportunity is no different,” he said.

What’s Included In The HPE-Nvidia GenAI Solution

HPE said its GenAI enterprise computing solution has been co-engineered with Nvidia and pre-configured to allow businesses of any size to use private data to quickly customize so-called foundation AI models, which have been trained on large data sets but can be adapted for a variety of purposes.

Businesses can then use these custom models as the foundation for GenAI applications that can run “anywhere,” from edge servers to cloud infrastructure.

“Ultimately, we feel like enterprises are either going to become AI-powered or they’re going to become obsolete, and the speed with which enterprises can embrace and explore and experiment with how they can transform parts of their business operations is really critical,” MacDonald said.

The joint solution starts with a rack-scale architecture consisting of 16 HPE ProLiant Compute DL380a servers, each of which have been pre-configured with Nvidia L40S GPUs, Nvidia BlueField-3 DPUs and the Nvidia Spectrum-X Ethernet networking platform.

HPE said the 16 servers have been configured with a total of 64 Nvidia L40S GPUs to fine-tune Meta’s popular and open-source 70-paramater Llama 2 foundation model.

On the software side, the solution features the HPE Machine Learning Development Environment software, which includes new GenAI capabilities for rapid model prototyping and tests, and the HPE Ezmeral software, which comes with new “GPU-aware” capabilities to aid with data preparation.

The solution also includes Nvidia AI Enterprise, the GPU designer’s suite of software that provides critical building blocks for bringing AI applications into production. This includes Nvidia’s NeMo framework, which includes methods for extracting proprietary data sets to use in GenAI models, fine-tuning those models and applying guardrails to ensure proper and safe use of those models.

Partners can order the solution for customers starting in the first quarter of next year.

New Solution Is Another Sign Of ‘Big AI Opportunity For Partners’

C.R. Howdyshell, CEO of Advizex, a Fulcrum IT Partners company which is in the process of attempting to close an AI deal in the range of $30 million to $40 million, said he sees the HPE-Nvidia enterprise GenAI solution as one more sign of the big AI opportunity for partners making the necessary investments to succeed in the AI solutions market.

“This is the future,” said Howdyshell of the AI opportunity ahead. “Advizex is heavily invested in providing these AI solutions. We have already invested in AI skill sets and are making more investments there. What’s more, our parent company Fulcrum IT Partners is looking to scale the AI business and acquire a partner with AI managed services and software expertise.”

The big opportunity for partners is around “educating’ customers on how they can leverage AI to drive significant competitive advantage, said Howdyshell. “The customers trust us and are counting on partners like us to help them through the AI solutions process,” he said.

Howdyshell said it is heartening to see the OEMs working closely with Nvidia to drive the AI solutions market forward. “That collaboration gives partners like Advizex a first mover advantage,” he said.

Critical to being successful in the AI market is a strong collaboration between Advizex, the OEMs and distributors, said Howdyshell. “We’ve been working on a $30-$40 million AI deal for five months,” he said. “It takes a concerted collective effort with the partner, OEM and the distributor to execute on these deals. There is a lot of work to get these deals done.”

Additional reporting by Steven Burke.