Nothing to Fear, Everything to Spin: Michael Burry Calls the AI Slowdown Push 'Self-Serving' IPO Hype

Two days after the CEOs of the world’s frontier AI labs united behind an unprecedented call to slow their own industry down, the most famous contrarian in American finance has offered a very different reading of their motives. Michael Burry — the investor whose bet against the mid-2000s housing bubble was chronicled in The Big Short — says the sudden wave of caution from OpenAI, Anthropic and other AI executives is not safety leadership at all. It is, in his words, “self-serving.”

In a late-Sunday post on X, followed by a longer piece on his Substack, Cassandra Unchained, Burry laid out a four-part indictment of the slowdown movement that has dominated the AI conversation since Anthropic CEO Dario Amodei published his 3,800-word “We Must Pace the Frontier” essay on September 12. “Let’s all take a moment to understand how self-serving it is for OpenAI, Anthropic and other execs of big hyperscalers to talk of slowing things down,” he wrote.

Burry’s four arguments

His case rests on four numbered claims, each more uncomfortable for the labs than the last:

1. LLMs are not AI and won’t be AGI. In Burry’s view, today’s large language models — including Claude and ChatGPT — will never achieve a “true capacity for reason,” so there is “nothing to fear” and “nothing AI to slow down.” This is the same argument Yann LeCun has been making for years, now weaponized against the safety push: if the technology isn’t actually on a path to superintelligence, the apocalyptic framing is theater.

2. A slowdown benefits incumbents. “Competition is coming up fast, slowing benefits incumbents,” Burry argued. This cuts at the heart of the labs’ credibility: OpenAI and Anthropic sit on the largest compute footprints and the deepest talent pools in the industry. A gentlemen’s agreement to ease off the accelerator — especially one with third-party “embedded evaluators” policing the pace — raises the cost of catching up for every smaller rival. What looks like restraint to a safety researcher looks like a moat to a short-seller.

3. IPOs need hype and puffery. This is the sentence that traveled furthest: “IPOs need hype & puffery; ‘we are so awesome it could become dangerous’ is hype & puffery.” Burry’s Substack post sharpened the point — Anthropic and OpenAI’s top brass, he wrote, have backed a slowdown “to increase the perception of their power before their IPOs.” The timing is hard to ignore: both companies have filed initial paperwork with the SEC, and both have watched their valuations climb toward the trillion-dollar zone on narratives of world-historical importance. If your moat is the story that your product might end the world, then warning about that power is not a risk to the story — it is the story.

4. Slowdown talk provides “cover for real uncontrollable slowing growth.” The most market-relevant claim: Burry suspects the industry’s growth is decelerating for reasons no one controls — saturation, rising inference costs, model plateau — and that a self-imposed pause gives executives a noble-sounding explanation for missing numbers. In his telling, the labs are pre-registering an excuse.

The context that made it land

Burry’s intervention did not happen in a vacuum. On September 12, Amodei published his essay urging the industry to voluntarily slow frontier development, proposing a three-part plan whose first step is “embedded evaluators” from third-party organizations like METR inside the labs. Sam Altman, Elon Musk and Demis Hassabis all publicly endorsed the call, an alignment of rivals with no real precedent.

The essay itself was catalyzed by insider dissent: former OpenAI and Anthropic researcher Jacob Coxon had accused both labs of “racing straight to self-improving superintelligence and gambling with our lives,” and Anthropic’s own Evan Hubinger said he puts the chance AI “kills all humans” at over 10% within a decade.

Then came the market reaction. On Monday, investors voted with their feet: Nvidia slid about 3% in premarket trading, while Intel, Micron and SK Hynix fell more than 5%, extending the selloff that began when the slowdown talk first hit the wires. President Trump weighed in over the weekend, blaming “very negative forces” for the apocalyptic chatter, insisting the worst “won’t happen,” and reiterating that the US must beat China in the AI race. Beijing’s own regulators formally rejected the US slowdown framing as a “Cold War playbook” — leaving the industry’s caution with few powerful friends.

The uncomfortable kernel

It is easy to dismiss Burry as a permabear with a book of shorts against the sector — he has spent 2026 warning about overinvestment in infrastructure, aggressive accounting, hidden debt and circular financing deals across the AI supply chain, and he writes publicly about his positions. OpenAI and Anthropic did not immediately respond to requests for comment on his claims.

But the reason his post spread beyond fintwit is that the incentive analysis is genuinely hard to refute. The labs warning about their own power is, functionally, a claim to power. A pause negotiated by the three best-funded players would bind their competitors more tightly than themselves. And both companies file IPO paperwork while their CEOs discuss extinction risk on national television — a combination that would trigger every fraud instinct Burry sharpened on mortgage CDOs.

The counterargument is equally real: Amodei’s proposal is not a training halt, and insiders like Coxon and Hubinger — people with no equity story to burnish — are the ones saying the risk is live. If they are right, dismissing all caution as marketing is the most dangerous spin of all.

Either way, the Burry salvo crystallizes the question the AI industry now faces in public: whether its leaders are uniquely honest about their product’s dangers, or uniquely skilled at monetizing the claim. The IPO windows — whenever they open — will force a verdict.

Sources

• Business Insider — Michael Burry says AI leaders’ calls for slowdown are ‘self-serving’
• Benzinga — Michael Burry Slams OpenAI, Anthropic AI Slowdown Calls as ‘Self-Serving’
• Yahoo Finance / GuruFocus — Michael Burry Sends Fresh Warning to OpenAI and Anthropic
• Reuters — Anthropic CEO urges AI companies to slow model development
• TechCrunch — Anthropic CEO outlines plan to pace the frontier