HPE AI solutions at Discover: Enterprise adoption insights

Artificial intelligence took center stage at HPE Discover last week. For 3 days, company executives captivated audiences with keynote sessions focused on Hewlett Packard Enterprise Co.’s cutting-edge AI solutions and the robust infrastructure it’s constructing to support them.

Latest solutions for AI deployment and the way HPE’s enterprise customers can benefit from them were discussed at length by theCUBE, SiliconANGLE Media’s livestreaming studio, through interviews with company executives, clients and industry analysts at HPE’s signature event. TheCUBE’s interviews included Antonio Neri (pictured), president and chief executive officer of HPE, who outlined the challenges for AI adoption during his keynote remarks. (* Disclosure below.)

“I assumed that was one in every of the strongest parts of Antonio’s keynote, where he said AI is tough and it’s filled with risks,” said Dave Vellante, chief analyst at theCUBE Research, in a post-keynote evaluation throughout the event. “He said they’ve been working on this five-point model framework. It’s got to be private. It’s got to be human focused. It’s got to be access for all. It’s got to be responsible. It’s got to be robust and tested in an ongoing way.”

Here’s theCUBE’s complete evaluation from Dave Vellante, who was joined by fellow analysts John Furrier and Rebecca Knight:

Take a look at these three essential insights you will have missed throughout the event:

1. HPE is leaning in on simplicity and accessibility for AI deployment.

Neri’s point about AI being hard has shaped his company’s work in developing AI solutions. The goal is to make sophisticated AI tools as accessible and manageable as possible for enterprise customers.

To this end, Neri announced HPE’s Private Cloud AI offering with Nvidia Corp.’s Chief Executive Jensen Huang throughout the opening day of the conference and emphasized a deal with minimal complexity and maximum efficiency.

“What we did yesterday with Jensen [Huang] is to deliver the amazing innovation that he has with amazing innovation HPE has in a single integrated experience and leverage that unified platform, called HPE GreenLake, to deploy [in] three clicks … 24 seconds,” said Neri, during his exclusive interview on theCUBE. “We’re going to point out up as one company selling the identical thing.”

Watch theCUBE’s complete video interview with Antonio Neri.

HPE’s Private Cloud AI solution was characterised by the corporate as a turnkey approach, an AI offering orchestrated to be seamlessly integrated into enterprise operations and tailored for businesses of all sizes.

“Our [AI solution is] already pre-integrated, pre-tested, and it’s able to roll within the 4 sizes that we have now and simply expandable,” said Jason Newton, vice chairman of integrated marketing, events and messaging at HPE, in conversation with theCUBE.

Here’s theCUBE’s complete video interview with Jason Newton, who was joined by Jim Jackson, chief marketing officer of HPE:

This message of simplicity provided HPE with a possibility to extol the virtues of GreenLake, its hybrid cloud platform designed to deploy and manage resources across private and non-private infrastructure. Private Cloud AI will probably be enabled by GreenLake, a part of HPE’s ongoing effort to assist customers centralize operations in a single platform.

“Since cloud adoption has began, the fact is that the companies, the enterprises have struggled to realize the cloud economics across the entire suite of their workloads,” said Tony Koinov, senior vice chairman and general manager of the GreenLake platform at HPE, during a discussion on theCUBE. “We have now built from the bottom up a platform-first cloud experience for them in order that we’re providing a single cloud, for his or her multivendor, multicloud estate … a single platform that rings common data, common services, common practices across all of their infrastructure.”

Here’s theCUBE’s complete video interview with Tony Koinov, who was joined by Latha Vishnubhotla, chief platform officer of HPE:

2. Nvidia is making its case for a distinguished role within the enterprise market.

The partnership between Nvidia and HPE that was unveiled throughout the Discover gathering in Las Vegas offered insight into the chipmaker’s enterprise interest.

Nvidia’s Bob Pette and HPE’s Neil MacDonald discuss their AI partnership.

Nvidia’s announcement of a brand new, microservices-based layer for its Nvidia AI Enterprise platform earlier this yr underscored its pursuit of the lucrative enterprise computing market. The enhancements included NIM microservices to enable optimized inference of popular generative AI models.

“We have now a slew of models covering quite a lot of enterprise use cases across multiple industries, which is what [CEO] Jensen refers to as the subsequent industrial revolution,” said Bob Pette, vice chairman and general manager of enterprise platforms at Nvidia, during an appearance on theCUBE. “We’re actually capable of take multiple data sources and predict, and that’s where increasingly more intelligence is extracted … these NIMs, these containerized models can come from anywhere. People can download them from wherever they’re, but we have now a nightly NIM AI factory at Nvidia that we proceed to construct these models, tune, test, optimize.”

The appliance of a containerized infrastructure for model processing is a glimpse into the longer term of AI architecture, where enterprises can integrate each data and model stacks for greater efficiency and price savings.

“The architecture just isn’t in regards to the individual nodes anymore. It’s not even in regards to the individual accelerators anymore,” said Neil MacDonald, executive vice chairman and general manager of the Compute Business Unit at HPE, who joined Pette in conversation on theCUBE. “It’s in regards to the whole system architecture that brings together all the capabilities in that compute stack, the information stack and the model stack that you might want to be effective with generative AI. So, one other shift here, each for our partnership and our collaboration but in addition for our enterprise customers, is pondering at that system level in a much, way more profound and cohesive way.”

Watch theCUBE’s complete video interview with Bob Pette and Neil MacDonald.

As Nvidia builds a presence within the enterprise and other market sectors, the corporate is maximizing its capabilities to power cloud and on-premises environments. Because of this the partnership with HPE makes an awesome deal of sense, in response to Holger Mueller, principal analyst at Constellation Research Inc., during an AnalystANGLE segment on theCUBE.

“Should you take into consideration Nvidia, they don’t have enterprise expertise,” Mueller said. “That is all happening recently, and all of the interest coming in is from the cloud hyperscaler side — which doesn’t give them the enterprise expertise either. So, who’s partner for [them], who has a long-term enterprise history and an install base … who knows the right way to construct complex systems?”

Here’s theCUBE’s complete video interview with Holger Mueller:

3. Key enterprise players are retooling infrastructure to fulfill AI needs without having to rebuild from scratch.

While AI has experienced remarkable advances in a brief time frame, it quickly became apparent that systems infrastructure needed to catch up. Firms reminiscent of Intel Corp. are working with clients to seek out optimal ways to integrate AI into existing computing architectures.

Intel’s Brett Hannath talks with theCUBE about processors for AI.

“What we’re attempting to do is bring AI compute power into architectures people already understand, know and use today,” in response to Brett Hannath, corporate vice chairman and chief marketing officer of Intel, during a conversation with theCUBE. “Slightly than having to go retrain all these infrastructure teams that you simply talked about, you possibly can actually leverage the architectures that you may have today. We actually are only attempting to play that role within the AI space so it’s not only the high end with the individuals with the massive wallets.”

Watch theCUBE’s complete video interview with Brett Hannath.

A technique that Intel is searching for to assist customers transition to the optimal infrastructure is by providing elasticity through recent chiplet designs, reminiscent of P-core or E-core processors which are focused on large AI workloads.

“The chiplet technology means that you can intersperse and select the precise node for the precise technology,” said Ryan Tabrah, vice chairman and general manager of Xeon E-Core products at Intel, during an interview on theCUBE. “Workloads will drive the software. The software will drive how the information center is built. The info center construct will drive how the rack is built, which then we just must have the products there that unleash all of those customers’ progressive ideas to make use of AI for awesome things.”

Watch theCUBE’s complete video interview with Ryan Tabrah.

In a quest to seek out the optimal AI infrastructure, enterprises have begun to expand data center capability. Chief technology officers at many firms are actually saying they’re purchasing more data center space for the primary time in over 10 years, in response to Zeus Kerravala, founder and principal analyst at ZK Research, during an appearance on theCUBE. This might soon result in a pullback in spending

“You’ve got a number of conflicting forces, in attempting to lower your expenses, in attempting to be more sustainable,” Kerravala said. “But I’m buying more AI, and I’m buying more data center space, and something’s got to provide. I’m concerned we’re going to see that within the back half of the yr.”

Here’s theCUBE’s complete video interview with Zeus Kerravala:

To look at more of theCUBE’s coverage of HPE Discover, here’s our complete event video playlist:

https://www.youtube.com/watch?v=videoseries

(* Disclosure: TheCUBE is a paid media partner for HPE Discover. Neither HPE and Intel Corp., the first sponsors of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE