On the CUBE Pod: Midterms shake up the AI ​​market

The very poor results of last week’s midterms have motivated technology investors go into a tailspin.

With many successful candidates running on platforms that oppose data centers and call for greater restrictions on artificial intelligence, investors are expecting a slowdown in AI spending. Tech stocks fell with a total loss of Market cap: $2.5 trillionand the Nasdaq fell more than 5.5%.

“There was a bloodbath yesterday,” he said John Furrier (pictured), Executive Analyst for theCUBE Research. “It’s been brutal here and tech stocks have been reeling. 41 states had data centers on the ballot. They all lost or won depending on how you look at it. Now it’s so mainstream, it’s a political football. So you’re starting to see the reality of AI. It’s affecting everyone. Everyone is affected.”

On the latest episode of theCUBE PodFurrier and Dave Vellante (Pictured right), principal analyst at theCUBE Research, discussed the fallout from the midterm elections and whether or not AI companies will really suffer in the long run. They also examined who will own the “system of intelligence” that will control the company’s AI.

Midterms put AI investors on shaky ground

After the “bloodbath” last week, the question remains: Can enterprise AI pay for the boom? Despite the shockwaves of the midterm elections, researchers at theCUBE predict a longer path to monetization. A recent analysis found that 2029 is the year with the highest risk for the AI ​​bubble.

“There is a clock on data center monetization and funding and how long it takes to build data centers,” Vellante said. “We know that anyone who can sell GPUs is making money, but companies aren’t self-funding yet. I said that by 2026, which is this year, you’ll see clear signs of monetization in the enterprise. I think you’re seeing it, but you’re seeing it in tech companies.”

According to Vellante, tech companies are using AI more effectively, which should come as no surprise. He believes the tech titans are in a solid position while neoclouds take on more of the risk by adopting a new software model, although not all of them will make it in the long run.

Furrier believes a GPU-based AI economy will transform in the next few years, which could lead to surprising winners and losers.

“Super micro [Inc.]good numbers, [Advanced Micro Devices Inc., Arista Networks, Marvell Technology Inc.]” he noted. “My opinion is this: infrastructure is evolving beyond GPU-centric computing.” Networking, optics, memory, storage, energy efficiency. System-level architecture becomes a strategic advantage, and these bottlenecks are being solved by the companies I just mentioned. You saw the news that Marvell completed a $115 billion acquisition of Lumentum.”

The stock market crash shows that not everyone will survive the eventual slowdown of the AI ​​era. Public opinion on data centers is mixed to say the least, so companies need to gain a foothold in the evolving system of AI intelligence to remain competitive.

“Neoclouds like CoreWeave will do it,” Vellante said. “They’re the first movers and they have some quality advantages, but they’re going to have to go through some hiccups. Hyperscalers are like an ATM. They could shut down at any time… but right now it’s all about growth, it’s about footprint, it’s about market share.”

Predicting the winners of the AI ​​era

Who owns the “system of intelligence”? This layer, which will bridge the gap between intelligent agents and the data systems they rely on, appears to still be in play. Palantir Technologies Inc. has the ontology that could give it a head start, while Snowflake Inc. and Databricks Inc. are still developing the operating systems – although Vellante sees the latter as one of the best strategies.

“You need determinism to have this intelligence system,” Vellante said. “So I think it’s completely open. At the moment I think Ali Ghodsi [chief executive officer of Databricks] has the best view. He goes for the client interface, he goes for the system of intelligence, he goes for the governance catalog, he goes for the agent action layer, and he works with all frontier model vendors. We’ll see if they can do it.”

Frontier models are the “wild card,” as Vellante puts it, but right now all tech companies are trying to get a piece of the pie. Despite recent market changes, Furrier sees AI-native companies as the ultimate outcome of this current phase of growth.

“The AI-native company or companies will be the model,” Furrier said. “That’s where everyone wins. Compute vendors, hyperscalers, the data platforms on the right, and the application companies, which will essentially be agents. Those will be the profit concentration winners. The control point… it won’t be the models, it’ll be data and context, and the agents will be the proof point.”

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

Tom Lee, investor and entrepreneur
Dan Ives, technology analyst
Frank Slootman, former CEO of Snowflake Inc.
Ali Ghodsi, co-founder and CEO of Databricks
John Roese, global CTO and chief AI officer of Dell Technologies
Lisa Su, Chairman and CEO of AMD
Michael Dell, Chairman and CEO of Dell Technologies
George Gilbert, principal analyst at theCUBE Research
Geoffrey Moore, Director at Geoffrey Moore Consulting
Rob Hof, Editor in Chief at SiliconANGLE Media
Stu Miniman, senior director of market insights, hybrid platforms, at Red Hat

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

Photo: SiliconANGLE