D.A. Davidson’s Gil Luria warns CoreWeave equity could go to zero if AI spending slows

D.A. Davidson’s Gil Luria warns CoreWeave equity could go to zero if AI spending slows

The analyst flagged a gap between roughly 9% borrowing costs and thin returns as shares fell more than 15% and bond yields climbed to 15-16%

CoreWeave has spent the AI boom as one of its loudest beneficiaries. On October 14, 2026, D.A. Davidson analyst Gil Luria offered a much less flattering read: the company’s equity could be worth nothing if the AI spending cycle cools.

Markets took the hint, sending CoreWeave (NASDAQ: CRWV) shares down more than 15% the same day.

The math Luria doesn’t like

Luria’s warning comes down to a basic spread problem. By his account, CoreWeave borrows money at roughly 9%.

The returns it earns on that capital look far thinner. Luria estimated the company’s return on invested capital at around 1%, or 4-5% at best.

Debt is the other half of the problem. CoreWeave’s borrowings now sit close to $50 billion, according to the research findings.

Advertisement

Luria cast the company as a potential example of the sector’s excesses. He called it a possible “poster child for bad behavior” in the AI infrastructure trade, pointing to its heavy leverage.

“Poster child for bad behavior” — Gil Luria, D.A. Davidson

Despite the gloomy framing, Luria kept a Neutral rating on the stock. He flagged the company’s thin margins and dependence on debt financing as red flags, stopping short of a formal sell call.

Bond markets are saying the quiet part out loud

Equity investors were not the only ones reacting. CoreWeave’s bonds also sold off, pushing yields up to 15-16%.

Credit default swaps, essentially insurance contracts against a borrower failing to pay, point in the same direction. Pricing on those instruments indicates a high risk of default over a five-year window.

A huge backlog, but no profits yet

CoreWeave reported a revenue backlog of $104.2 billion as of the end of the second quarter of 2026.

Yet the company has never posted a profit. A $104.2 billion order book is reassuring only if the economics of filling those orders eventually turn positive.

CoreWeave went public in March 2025 with backing from Nvidia. Since then, it has had to manage significant operational risks, including steep interest expenses and execution challenges as it builds out capacity.

What this means for investors and the AI trade

The most direct stakes fall on CoreWeave shareholders and bondholders, and their interests are not identical. In a stress scenario, creditors stand ahead of equity holders in line. That is why Luria’s warning specifically targets equity value: if things go badly, stockholders are the first to absorb losses.

The key variable is AI spending itself. Luria’s zero scenario is conditional, tied to a slowdown in the AI investment cycle rather than presented as a base case.

Luria’s estimate of returns between roughly 1% and 4-5%, set against borrowing costs of about 9%, is the number to keep in mind. Until that gap closes, the debate over CoreWeave is less about whether demand exists and more about whether the business can ever earn more on its capital than it pays to borrow it.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.
D.A. Davidson’s Gil Luria warns CoreWeave equity could go to zero if AI spending slows
D.A. Davidson’s Gil Luria warns CoreWeave equity could go to zero if AI spending slows

The analyst flagged a gap between roughly 9% borrowing costs and thin returns as shares fell more than 15% and bond yields climbed to 15-16%

CoreWeave has spent the AI boom as one of its loudest beneficiaries. On October 14, 2026, D.A. Davidson analyst Gil Luria offered a much less flattering read: the company’s equity could be worth nothing if the AI spending cycle cools.

Markets took the hint, sending CoreWeave (NASDAQ: CRWV) shares down more than 15% the same day.

The math Luria doesn’t like

Luria’s warning comes down to a basic spread problem. By his account, CoreWeave borrows money at roughly 9%.

The returns it earns on that capital look far thinner. Luria estimated the company’s return on invested capital at around 1%, or 4-5% at best.

Debt is the other half of the problem. CoreWeave’s borrowings now sit close to $50 billion, according to the research findings.

Advertisement

Luria cast the company as a potential example of the sector’s excesses. He called it a possible “poster child for bad behavior” in the AI infrastructure trade, pointing to its heavy leverage.

“Poster child for bad behavior” — Gil Luria, D.A. Davidson

Despite the gloomy framing, Luria kept a Neutral rating on the stock. He flagged the company’s thin margins and dependence on debt financing as red flags, stopping short of a formal sell call.

Bond markets are saying the quiet part out loud

Equity investors were not the only ones reacting. CoreWeave’s bonds also sold off, pushing yields up to 15-16%.

Credit default swaps, essentially insurance contracts against a borrower failing to pay, point in the same direction. Pricing on those instruments indicates a high risk of default over a five-year window.

A huge backlog, but no profits yet

CoreWeave reported a revenue backlog of $104.2 billion as of the end of the second quarter of 2026.

Yet the company has never posted a profit. A $104.2 billion order book is reassuring only if the economics of filling those orders eventually turn positive.

CoreWeave went public in March 2025 with backing from Nvidia. Since then, it has had to manage significant operational risks, including steep interest expenses and execution challenges as it builds out capacity.

What this means for investors and the AI trade

The most direct stakes fall on CoreWeave shareholders and bondholders, and their interests are not identical. In a stress scenario, creditors stand ahead of equity holders in line. That is why Luria’s warning specifically targets equity value: if things go badly, stockholders are the first to absorb losses.

The key variable is AI spending itself. Luria’s zero scenario is conditional, tied to a slowdown in the AI investment cycle rather than presented as a base case.

Luria’s estimate of returns between roughly 1% and 4-5%, set against borrowing costs of about 9%, is the number to keep in mind. Until that gap closes, the debate over CoreWeave is less about whether demand exists and more about whether the business can ever earn more on its capital than it pays to borrow it.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.