Michael Burry says AI is a 'byzantine' money loop — Nvidia's monster results could be ammo for his argument

On the earnings call, CEO Jensen Huang proudly proclaimed, “We have a golden age of new AI labs and startups, multiple frontier labs scaling in parallel, a thriving open-model ecosystem and physical AI coming online — with strong momentum across the U.S. and around the world.”

Judging by the roughly 8% pop for Nvidia shares the next day, it’s clear many investors believe the future is rosy for the AI growth story.

But “Big Short” investor Michael Burry isn’t a buyer. If anything, Nvidia’s sterling report could be used as further evidence of his long-term bearish thesis on overinflated AI demand.

Before the earnings release, Burry pointed out the seeming disconnect between the AI growth narrative and Nvidia’s year-to-date performance, writing on Substack, “Nvidia’s current stock price is not congruent with the market’s narrative. On the face of it, the stock is wildly undervalued. Low PE for a big grower that currently commands monopoly rents. Yet, NVDA stock is treading water compared to recent years’ performance.”

One of Burry’s long-standing arguments for shorting AI stocks like Nvidia is that their demand is more about circular financing than genuine sales to customers. As The Street reported, Burry characterized these deals as “byzantine financing arrangements” that create a distorted view of the data center boom.

Compute needs Nvidia’s capital

On August 11, Nvidia tackled these accusations of “circular financing” head-on with a fresh arrangement to fuel AI’s next stage of growth. In this announcement, Nvidia said it was partnering with six financial institutions — including BlackRock, Apollo and Goldman Sachs — as “independent financing platforms” ready to inject $500 billion into AI infrastructure over the coming years.

Nvidia deliberately said this new arrangement should quell fears of circular financing in AI, arguing, “The capital providers independently underwrite each project — including the customer, demand, utilization, cash flow and residual value. Nvidia provides the platform; the investors make independent financing decisions.”

They also repeatedly assured readers that “the demand is real” for its GPUs, citing interest from “frontier AI labs, AI-native startups, enterprises, cloud providers and countries building AI services.”

In reality, this news only pushed some financial analysts further toward Burry’s cynical stance on chip sales.

One potentially worrying part of this whole arrangement is that Nvidia says it may guarantee up to 25% of the value of some projects — or, as they put it in their blog post, provide a “residual-value support mechanism.” So, even though Nvidia isn’t footing the entire bill, it does have a sizable financial stake in helping some AI projects get funded.

In a report in the Financial Times, Morgan Stanley estimated Nvidia’s total potential credit exposure could reach$200 billion by 2028 as it becomes more integral in financing AI infrastructure. As Morgan Stanley’s analysts warned, “Conventional leverage increasingly understates the [Nvidia] credit story as ecosystem support sits in contingent, contractual, and potentially off-balance-sheet forms.”

This credit risk exposure is a reason why Morgan Stanley rated Nvidia’s stock as “neutral” rather than a “buy.”

D.A. Davidson’s Head of Technology Research, Gil Luria, raised a similar fear about how durable AI demand is, given Nvidia’s financer status. In a New York Times interview, Luria questioned, “How big would the market be if Nvidia weren’t funding its customers?”