Michael Burry reveals his verdict on the ongoing AI bubble

16 Sep 2026, 19:17 · The Fresno Bee

Every time Silicon Valley’s biggest names line up behind a single message, someone eventually asks who benefits from such universal agreement. This time, that someone was the investor best known for calling the housing crash before almost anyone else saw it coming.

Michael Burry has spent the past year building a reputation as one of the AI industry’s loudest skeptics. His latest target is not a stock but a story. When three of the sector’s most powerful executives suddenly agreed the technology needed to slow down, Burry saw the timing as less about caution and more like a self-serving pitch.

Michael Burry calls the AI slowdown self-serving

Burry has spent much of 2026 building short positions against companies tied to the AI trade, disclosing bets against Nvidia, Tesla, Micron, Applied Materials, Caterpillar and a leading semiconductor ETF, according to TheStreet.

On September 14, Burry published a post on X and on his Substack, Cassandra Unchained, arguing that people should take a moment to understand how self-serving it is for OpenAI, Anthropic and other big hyperscaler executives to talk about slowing things down. The post circulated quickly across financial media, Yahoo Finance reported.

The post arrived days after a wave of public agreement among rival AI labs. Anthropic chief executive Dario Amodei published an essay calling for a deliberate slowdown in AI capability development. The idea was quickly supported by OpenAI’s Sam Altman, Elon Musk and Alphabet’s Demis Hassabis. A rare moment of alignment among executives who normally compete fiercely for compute and talent.

Burry’s broader argument has centered on a feedback loop in the AI trade: the idea that chip stocks rise because hyperscalers spend heavily, partly because chip stocks keep rising. He has argued that a feedback loop can look like real demand even when it is partly self-reinforcing hype.

Inside Anthropic CEO’s slowdown manifesto

Amodei’s roughly 3,800-word essay, titled “We Must Pace the Frontier,” argued companies should independently monitor model development, coordinate with other democratic AI labs, and eventually reach an agreement with authoritarian governments including China. He framed the plan as pacing rather than halting AI progress, CNBC reported.

Altman was on board within hours. So was Musk, who posted three words: “Dario is right.” Demis Hassabis of Google DeepMind followed. Three labs that normally fight over every GPU and every talent hire suddenly had the same public position.

The China provision proved the most contentious part. Amodei urged Washington to maintain tightening chip export controls and take stronger action against unauthorized model distillation, while acknowledging that China’s willingness to cooperate remained the toughest unresolved question. Beijing’s state-run Global Times dismissed the essay as a Cold War playbook. President Trump separately stated the U.S. was leading China in AI and had no interest in slowing down, according to CNBC.

Bloomberg / Getty Images

Burry’s four reasons for his skepticism

Burry laid out his skepticism in four numbered points. His first objection challenges the entire premise: he argues that large language models are not artificial intelligence and will never become artificial general intelligence. “LLMs are not AI and won’t be AGI. There is nothing AI to slow down,” he wrote.

His second point turns to competition. Burry argues that slowing frontier development benefits whichever companies are already ahead, since fast-moving rivals lose the most ground when the pace of the whole industry deliberately eases.

His third reason zeroes in on timing. Both OpenAI and Anthropic filed confidential paperwork for initial public offerings earlier this summer. OpenAI’s Sam Altman has said an IPO this year would be an “ill-advised moment,” while Anthropic is targeting an October listing on the Nasdaq at a valuation that could reach approximately $2 trillion. Burry argues that warnings framed as “we are so awesome it could become dangerous” function as marketing ahead of some of the largest listings Wall Street has ever seen, Yahoo Finance reported.

His fourth point is the sharpest. Burry suggests the safety talk could provide cover for a slowdown in growth that has nothing to do with caution, pointing to Altman’s own admission that an OpenAI IPO this year would be ill-advised. Safety concerns and IPO incentives may be more closely intertwined than the companies acknowledge, he argued.

Other skeptics and the bigger picture

Burry is not alone in questioning the motives behind the push. David Sacks, the White House AI and crypto czar, argued that if Anthropic and OpenAI genuinely believed their models were too dangerous to release, they could simply slow down on their own rather than tying that restraint to new regulatory frameworks first.

Anthropic’s own house is not fully in agreement either. The company’s head of economics put out data this summer showing no clear AI-driven job losses in the broader economy. That sits awkwardly next to Amodei’s warnings of a looming white-collar bloodbath.

The IPO filings are in. The roadshows are coming. That is when the safety language gets stress-tested against actual financials. Burry’s reading is simple: watch the timelines, not the essays.

The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.

This story was originally published September 16, 2026 at 12:17 PM.