Nvidia Cast in Co-Starring Role at HPE Discover 2024

HPE DISCOVER — HPE on Tuesday used its 2024 user conference in Las Vegas to position itself as best buddies with Nvidia, which has no shortage of artificial intelligence (AI) friends these days.

The companies launched Nvidia AI Computing by HPE, a portfolio of co-developed products and services designed around generative AI.

HPE rolled out the first part of the portfolio at Discover. HPE Private Cloud AI integrates Nvidia Spectrum-X Ethernet networking, HPE GreenLake for File Storage and HPE ProLiant servers with support for Nvidia L40S and H100 NVL Tensor Core GPUs, and the Nvidia GH200 NVL2 platform. The Private Cloud AI software stack includes the Nvidia AI Enterprise storage platform with NIM inference microservices and HPE AI Essentials. The AI software provides curated AI and data foundation tools with a unified control plane.

The HPE Private Cloud AI is available through the HPE GreenLake cloud. It includes HPE’s new OpsRamp AI copilot to increase workload and IT efficiency.

Nvidia AI Computing by HPE offerings and services will be available through a joint go-to-market strategy that spans sales teams and channel partners, training, and system integrators including Deloitte, HCLTech, Infosys, TCS and Wipro.

HPE also launched an AI enablement program with Nvidia that includes new competencies and workshops tied to both vendors’ technologies, products and services.

Related:HPE Discover 2024: HPE Virtualization Draws VMware Comparison

HPE is hardly the only IT player partnering with Nvidia on generative AI. A big part of its AI strategy is to convince partners and customers that its technology works better with Nvidia than its rivals' work with the GPU giant. Nvidia played a large role at HPE Discover. Its CEO, Jensen Huang, appeared alongside HPE CEO Antonio Neri at the Tuesday keynote session at the Sphere, and the conference agenda was packed with other Nvidia speakers.

Huang said to build generative AI, “on a practical basis, you need a model stack, a data stack and a computing stack. Each one of those stacks is incredibly complicated. Between our two companies we’ve turned these three stacks into deployable solutions.”

Nvidia NIMs make up the model stack, Nvidia AI Enterprise is the data stack and HPE provides the computing stack and software that go along with that.

"We want to take those three stacks and simplify it into a private cloud," Huang said.

HPE Discover 2024 Features Turnkey Private Cloud AI

“At HPE, we wanted to create something unique with Nvidia,” Neri said. “Unlike services-led AI solutions out there, we deliver this as one integrated product,” Neri said. “It’s ready to run out of the box.”

HPE Private Cloud AI comes in four configurations — small, medium, large and extra large. Customers can start with any of the four and scale up to a larger configuration.

Fidelma Russo, HPE CTO and EVP of hybrid cloud, said what sets Private Cloud AI apart from rivals is it is a turnkey system rather than a reference architecture.

HPE's Fidelma Russo

“It’s ready to run out of the box,” she said. “You plug it in, you connect it to the GreenLake cloud, and three clicks later − boom − your data science and IT operations teams are up and running the Nvidia software stack.”

HPE Private Cloud AI includes:

HPE GreenLake for File Storage has achieved NVIDIA DGX BasePOD certification and Nvidia OVXstorage validation, which HPE executives say make it a proven enterprise file storage solution for running AI, GenAI and GPU-intensive workloads at scale. HPE also is planning Nvidia reference architecture storage certification programs.

HPE GreenLake for File Storage uses Vast Data technology through an OEM deal. File data is crucial for running generative AI workloads.

“Every discussion with our customers starts around data, what they're going to do with their data, and that now turns into a discussion around AI,” Russo said. “And how they're going to transform their business and use new data.”

Expect Private Cloud AI to be generally available in the fall. HPE also expects to ship Cray and ProLiant servers with Nvidia GPUs this year.

Look for HPE Cray XD670 servers with NVIDIA H200 NVL this summer, with ProLiant DL380a Gen12 servers with Nvidia H200 NVL Tensor Core GPUs and ProLiant DL384 Gen12 servers with dual Nvidia GH200 NVL2 due in the fall.

HPE’s AI partner strategy with Nvidia includes trainings and certifications such as AI acceleration workshops, AI skills development workshops, self-paced content and certifications; and competency programs for HPE AI solutions, compute, high performance computing (HPC) for enterprise, and storage and data.