Oracle is not an enterprise AI bellwether

Welcome to Runtime! Today: Wall Street dreams that Oracle could be the new face of enterprise AI will be deferred, Qualcomm hedges its Arm server chip bets, and the latest enterprise moves.

(Please forward this email to a friend or colleague! And if it was forwarded to you, sign up here to get Runtime each week.)

Red-faced

For a company that helped invent the concept of enterprise information technology as much as any other, Oracle has been remarkably late to just about every major shift in this market over the course of the 21st century. One thing it is very good at, however, is attracting an outsized amount of attention compared to its actual impact.

Oracle's stock fell more than 10% Thursday after it reported second-quarter earnings that missed Wall Street's revenue expectations and revealed that it is burning $10 billion in cash a quarter on new infrastructure, setting off new fears about the collapse of the AI bubble. A year that started with Oracle co-founder Larry Ellison sharing an Oval Office podium with OpenAI CEO Sam Altman and cash pyromaniac Masayoshi Son is ending with widespread concerns that Oracle has bitten off more than it can chew.

• Oracle's cloud infrastructure revenue during the last quarter was just $4.1 billion, which is up an impressive 68% compared to last year but isn't much more than what AWS brings in during a given week.
• But Oracle reported that "total remaining performance obligations" were up 438% year-over-year to $523 billion.
• The company's problem is that nobody has any idea how it will get from doing roughly $16 billion a year in infrastructure revenue this year to roughly $300 billion over the next few years on the backs of a handful of customers.

The backbone of those "total remaining performance obligations" — the massive computing deal Oracle signed with OpenAI in September — now looks like a fever dream after it was considered a breakthrough moment for its cloud infrastructure ambitions by Wall Street. There are two main reasons for the increased skepticism, and Oracle has limited control over both of those problems.

• The first problem for Oracle is that OpenAI needs to generate an incredible amount of revenue over the next couple of years to pay for those infrastructure spending commitments, and its path to doing so is murky at best.
• The second is that unlike its hyperscaler rivals, Oracle is relying heavily on debt to support its capex expansion, which in past tech bubbles has been an enormous red flag should demand start to wane.
• "While heightened capex and debt expenses, as well as the tight alignment with OpenAI's success, make this somewhat of a show-me story, we believe that Oracle should be a major beneficiary of the AI platform shift, as it builds critical computing capacity for a growing set of hyperscaler, AI labs, and enterprise customers," said William Blair analyst Sebastien Naji, according to Yahoo Finance.

For those reasons, anyone trying to understand how enterprise AI spending will evolve in 2026 probably shouldn't look to Oracle's results, especially given the widespread reluctance among a generation of CIOs and enterprise tech leaders to do business with a company known for its scorched-earth business practices, Oracle may well figure out how to pull off the infrastructure scaling event of all time, and OpenAI might find a way to break through to the enterprise, but they are more intertwined than representative of the market in general.

• The Big Three seem quite pleased with the momentum of enterprise AI over the last year, and even Salesforce reported better-than-expected progress after getting out way over its skis earlier this year.
• And the tools are starting to come together that could help enterprises fulfill the promise of agentic AI, especially now that the industry has agreed to support open standards for agent connectivity.
• The Next Platform agreed with Naji that Oracle is laying a foundation for the next decade of enterprise infrastructure, but as he put it, right now Oracle is "a show-me story."

Looking for a way to support independent tech journalism this holiday season? For $10 a month, you'll help us continue our mission to bring reliable and actionable coverage of this vital sector of the economy and gain access to supporter-only features currently in the works, such as an exclusive discussion and events forum.

RISC it for the biscuit

Qualcomm has been flirting with the enterprise server chip market for almost a decade, announcing plans in 2017 to produce Centriq server processors based around Arm's chip designs only to abruptly shutter the division around a year later. Last month the company announced that it was getting into the enterprise GPU business with new Arm designs, and Wednesday it announced that it had acquired Ventana Micro Systems, a startup working on CPUs designed around the RISC-V instruction set.

RISC-V is an open-source chip instruction set architecture, which is basically a set of marching orders governing how chips should execute the instructions passed down to them by software. It has attracted a lot of attention over the last several years thanks in large part to the fact that it's a royalty-free way for chip makers and designers to get up and running, but RISC-V chips have not seen a lot of adoption.

Ventana could allow Qualcomm to compete with Nvidia by offering an alternative package of GPUs and CPUs as part of a complete server product akin to Nvidia's Grace Hopper product, but that would be a sizable undertaking. The data-center chip market remains in flux as Intel scrambles to find its footing, but Qualcomm would be starting from scratch as AMD and homegrown Arm chips from hyperscalers like AWS's Graviton see increased adoption.

Enterprise moves

Chano Fernández is the new co-CEO of Klayvio, with plans to focus on sales and operations alongside co-founder and co-CEO Andrew Bialecki.

Michael Gordon is the new chief operating officer and chief financial officer at Crusoe, joining the neocloud after serving in the same role at MongoDB for a decade

Denise Dresser is the new chief revenue officer at OpenAI, joining the company after serving as CEO of Salesforce's Slack with a mandate to focus on the enterprise.

Derek Phillips is the new chief revenue officer at Vectra AI, joining the cybersecurity company after leading sales at Claroty.

Heather Planishek is the new chief financial officer at Lambda, joining the one of the other prominent neoclouds after operations and financial leadership roles at Tines and Palantir.

Amin Vahdat is the new chief of AI infrastructure at Google, according to Semafor, a promotion reporting directly to CEO Sundar Pichai after nearly 15 years of working on infrastructure at the company.

The Runtime roundup

Broadcom reported revenue and earnings that beat Wall Street expectations and raised guidance for the current quarter, declaring that sales of its AI chips would double compared to last year.

Microsoft announced that it would extend its bug bounty program to reward the discovery of "critical vulnerability has a direct and demonstrable impact to our online services," including open-source software.

Thanks for reading — see you Saturday!