Furrier: AI backlash an "AI freakout moment" as IBM bets on commodity models

The podcast conversation lands against a rough stretch for AI-linked stocks. Last week shares of Nvidia, CoreWeave, Oracle and Nebius declined amid reports that OpenAI's annualized net revenue was approximately $50 billion, roughly $20 billion below previously reported figures. Furrier's response is that the public uproar is a "freakout moment" rather than a fundamental break, pointing to what he says is happening inside the supply chain, where the data center is now treated as a system and rack scale is changing.

On IBM, Vellante's read from the company's recent Analyst Council is that customers are spending first on the infrastructure and models needed to produce intelligence, and only then turning to IBM to put that intelligence to work. That sequencing is the challenge as the AI market matures, he suggested. IBM has promoted a four-part architecture of agents, data, automation and hybrid infrastructure with an orchestration layer that can integrate agents from different vendors, and Furrier framed the company's core bet as the commoditization of models, adding that IBM doesn't need its own proprietary operating system.

In chips, Marvell is drawing attention with an ambitious revenue target of $70 billion to $90 billion for fiscal 2031, and earlier this year announced a partnership with Nvidia's NVLink Fusion. Furrier said what impressed him was Marvell's 10-year vision, including an argument from a decade ago that AI research is becoming a connectivity problem as much as a compute problem. Neither host sees an imminent bubble: Furrier doesn't believe the market is approaching one, and Vellante sees the breaking point as farther away than many might expect, as enterprise focus shifts from training to inference.