CoreWeave and Oracle Debt: Are AI Debt Fears Overdone?

The short answer: CoreWeave and Oracle debt has become the market's biggest AI worry, with credit-default-swap spreads on both widening sharply in October. Dan Ives of Yorkville Ives agrees that rising rates raise the cost of capital. But he argues CoreWeave's debt is small relative to its contracted backlog and that AI demand still exceeds supply. He says he'd "bet against the negative narrative."

From The Pomp Podcast, hosted by Anthony Pompliano, with guest Dan Ives, partner at Yorkville Ives and one of Wall Street's most-followed tech analysts. Episode released October 22, 2026.

CoreWeave and Oracle debt is at the center of the AI sell-off. Both companies have borrowed heavily to build data centers, and in October investors started demanding a lot more to insure that debt. On the October 22 episode of The Pomp Podcast, host Anthony Pompliano asked Dan Ives, one of Wall Street's most-followed tech analysts, what the bears have right, what they have wrong, and what could actually slow AI demand.

The market data is stark. According to CNBC's analysis of FactSet data, CoreWeave's five-year credit default swap spread reached about 1,066 basis points on October 21. Oracle's rose to about 308 basis points, up from 141 at the start of the year. A CDS spread is the cost of insuring against default, so wider spreads mean investors see more risk.

CNBC also reported that CoreWeave has taken on more than $51 billion of debt and obligations to fund its build-out, and that an S&P Global analyst described spreads above 1,000 basis points as generally a sign of credit distress. Per the same CNBC reporting, S&P downgraded Oracle's debt to the lowest investment-grade rating in July.

Ives called Oracle and CoreWeave the "poster children" of companies taking on a lot of debt to build out AI capacity.

Ives didn't dismiss the concern. He said there's "no denying" that rising yields put more pressure on the AI trade. He estimates roughly 20% of the AI buildout will need to be funded with debt or some kind of equity or convertible financing. When rates rise, that capital gets more expensive.

"So they're right, I think, directionally," he said of the bears' point on financing costs. He pointed to CoreWeave as the example.

He also drew a clear line between borrowers. Hyperscalers like Microsoft, Amazon and Alphabet are in a much stronger position thanks to their cash and existing businesses. Neoclouds like CoreWeave don't have that cushion. For them, he said, heavy borrowing is a strategic choice they've had to make.

Ives' main counterargument is about scale. In his view, even for CoreWeave, "the debt that they're taking on relative to their backlog is minuscule." He said Nvidia CEO Jensen Huang has made a similar point about CoreWeave's health. (We couldn't confirm Huang's exact wording, so treat that as Ives' characterization.)

He compared the moment to the "SaaS apocalypse" scare in software earlier this year: a narrative that, in his view, ran far ahead of fundamentals. He calls this a SaaS-apocalypse-style narrative moment for AI infrastructure.

And he thinks the world needs more neoclouds, not fewer, because demand will outrun supply. Huang separately told CNBC that neocloud demand is "off the charts" while warning of a "scarcity of capital," per the same CNBC report. Ives' bottom line: "I'll bet against the negative narrative here."

Pompliano noted that the Fed had raised rates and the stock market went higher anyway. Ives' read is that rates are rising partly because of the AI capex boom itself, which he calls a fourth industrial revolution funded by U.S. tech.

He acknowledged another hike could come. But unless the 10-year Treasury yield surges dramatically, he doesn't think higher rates will significantly slow the buildout. In his view, they mostly create jitters, which show up in CoreWeave's CDS and Oracle's stock.

Pompliano argued he can't see what would slow demand. Every new AI product release seems to accelerate usage, and inside his own companies, AI is used for video editing, coding, growth analytics and SEO. He said people who six months ago used AI as a Google replacement are now diving in head first.

Ives agreed. He said consumer AI is going mainstream, physical AI hasn't started, and the cycle is less than 15% complete. We cover that view in Dan Ives AI boom call: 85% is still ahead of us. The risks he did flag were rates, a political pushback on data centers around the midterms, and companies that lean too hard on borrowing.

Both borrowed heavily to build AI data centers. Rising rates and the cost of insuring their debt, measured by CDS spreads, jumped in October 2026.

He says its debt is minuscule relative to its contracted backlog and calls the fear a narrative-driven moment, though he agrees higher rates raise capital costs.

Ives doesn't think so unless yields surge dramatically. He expects rate hikes to create jitters rather than significantly slow the buildout.

According to Ives, yes. Hyperscalers fund AI from large cash flows, while neoclouds rely more on debt.

Anthony Pompliano is an entrepreneur, investor, and author who serves as Chairman and CEO of Silvia, Inc (SVIA) and invests through Pomp Investments. He is the host of The Pomp Podcast, where he interviews founders, investors and economists about bitcoin, markets, AI and the economy. This article is based on his October 22, 2026 conversation with Dan Ives, with outside facts checked against the sources listed above. Follow Pomp on X (https://x.com/APompliano).

Get notified for every new Pomp Podcast episode.

Subscribe for free to learn alongside Anthony Pompliano.

This article reflects the opinions of the guest and host as shared on The Pomp Podcast. It is guest opinion, not investment advice. Do your own research before making any investment decision.