On the CUBE Pod: IBM competes on AI’s operating model

American have major concerns about artificial intelligencebut that has OpenAI Inc., Nvidia Corp., IBM Corp. and not stopping others from moving forward.

Last week saw shares of Nvidia, CoreWeave Inc., Oracle Corp. and Nebius Group NV declined after reporting that OpenAI’s annual net revenue was about $50 billion, about $20 billion lower than previously reported numbers. Despite the ongoing public resistance against AI, John Furrier (Pictured right), Research Executive Analyst at theCUBE Research, believes the upheaval is temporary.

“This is an AI freakout moment,” he said. “It’s funny to watch when you know what’s happening on the inside. We’re talking to all these suppliers, be it Nvidia, the chip side… the Neoclouds or the AI ​​infrastructure clouds. This new AI layer is a system; the data center as a system is now the computer. The rack size is changing. Companies like Marvell, Broadcom and others are developing chips to achieve greater efficiency.”

On the latest episode of theCUBE PodFurrier and Dave Vellante (Pictured left), principal analyst at theCUBE Research, discussed insights from the recent IBM Analyst Council. They also examined the growth of chipmaker Marvell Technology Inc. and the industry’s shift toward inference.

IBM relies on commodity models

Big Blue’s most recent meeting offered insights into the traditional company’s position in the AI ​​age. Vellante says customers are increasingly spending money on the infrastructure and models needed to produce information before turning to IBM to put that information into action. This presents a challenge for the company as the AI ​​market matures.

“The alpha will be in your data, your processes and your operating model,” Vellante said. “That is [IBM’s] Opportunity. This AI operating model will be the new source of competitive advantage. It won’t come from Palantir; It won’t come from models. It depends on your own proprietary data, how you apply that data, how you optimize your processes, and how quickly you can bring products and services to market.”

IBM sponsored a four-part architecture consisting of agents, data, automation and hybrid infrastructure, with an orchestration layer that can integrate agents from different providers. Rather than relying solely on proprietary models, IBM is positioning itself as a provider of hybrid AI operational infrastructure for enterprise customers, building on Red Hat’s open source foundation.

“Their goal is to commercialize the models,” Furrier said. “Having a good model management capability will be good for IBM, but if you’re going for commercialization, you better be right. IBM doesn’t need its own proprietary operating system.” [because] … they have a valuable abstraction control layer around the open infrastructure through Red Hat. Red Hat gets paid a lot of money to manage this open abstraction. This is an opportunity: IBM could run all the agents.”

The question remains whether IBM’s commitment to commodity models and open source solutions will pay off. Customers trust IBM. However, according to Vellante, they do not see the company as an innovator in the current AI landscape. However, if IBM can establish itself as a provider Operating infrastructure around AI models and agents, then it could have “the scarcity around the abundance”.

“You have to play in both camps,” Vellante added. “In other words, they must have IBM’s version of Oracle’s Red Stack. We asked last year [IBM’s CCO] Rob Thomas, what is the platform? And the answer came back: Red Hat. I agree. Let’s build a solution around Red Hat to make AI adoption easier. It’s unclear to me what this solution is. It’s clear to me that they have all the pieces.”

The AI ​​market is entering the inference phase

In the chip world, Marvell Technology Inc. is drawing attention with an ambitious sales target of $70 billion to $90 billion for the 2031 fiscal year. Earlier this year, the company, which makes custom AI processors and optical networking technology, announced a partnership with Nvidia’s NVLink Fusion.

“What impressed me about Marvell is that they had a 10-year vision,” Furrier said. “Their mission statement that they put out 10 years ago… they didn’t have factories. That wasn’t part of the dialogue. But back then… they made the argument that AI research was becoming as much a connectivity problem as it was a computing problem. And they’re not wrong. There’s a semiconductor company that’s talking about economics. They’re talking about real value.”

Furrier doesn’t believe the AI ​​market is nearing a bubble, while Vellante sees the breaking point as far away than many might expect. As the corporate world shifts its focus from training to inference, companies like Marvell and the Neoclouds will benefit financially as AI infrastructure spending increases.

“A year ago, 90% of workloads – this is what we saw at CoreWeave – were training,” Vellante said. “Today it’s like 60-40 training-inference. It’s expected to flip to 10-90. Why is that? Because they increase sales; they make money from inference. They talk about it all the time. You go to college to get an education, and then you make money and you’re constantly learning.”

Check out the full podcast below to find out why these industry pros were mentioned:

Ray Dalio, chief investment officer of Bridgewater Associates
Matt Calkins, CEO of Appian
David Lazovsky, EVP and GM of the Data Center Networking Business Group at Marvell Technology
Arvind Krishna, Chairman and CEO of IBM
Rob Thomas, SVP of Software and Chief Commercial Officer at IBM
John F. Akers, former CEO of IBM
Prith Banerjee, SVP of Innovation and Executive Team Member at Synopsys
Matt Murphy, Partner at Menlo Ventures
Charlie Kawwas, President of Broadcom
Donald Trump, 47th President of the United States of America

Here is the full episode of this week’s theCUBE Pod:

You can also listen to the CUBE Pod Apple Podcasts or further Spotify. For video episodes visit us YouTube playlist. Tune in now and join the ongoing discussion.

Photo: SiliconANGLE