One of the key players who predicted the 2008 financial crisis has come out swinging against the latest, co-ordinated doomsday chorus emanating from Silicon Valley.
Michael Burry, who made a fortune from predicting the US housing market crash almost 20 years ago, says there may be an ulterior motive inside the AI industry’s sudden call-to-arms.
He believes the latest panic is a ploy to frighten the public, slow the competition and keep the biggest AI companies looking indispensable as they prepare to sell themselves to Wall Street.
It’s a brutal theory that is landing at an unusually turbulent moment for the planet in its delicate financial state.
Want more AI news? Download our new app for even more coverage
Even those developing and marketing the world’s most powerful systems have begun sounding like panicked techies sprinting out of the engine room.
The past few weeks have been quite illuminating indeed.
OpenAI chief scientist Jakub Pachocki went out and wrote an essay saying nobody is prepared for the consequences of rapidly rising machine intelligence. And then former OpenAI and Anthropic researcher Jacob Coxon quit the game in a huff and accused both companies of “gambling with our lives”.
Then Anthropic boss Dario Amodei stressed that frontier development should be slowed immediately through external watchdogs.
Sam Altman agreed.
Elon Musk also agreed.
Then we have the other billionaires who say the panic is overblown and that the West’s relentless pursuit of superintelligence is our most important goal, with rivals like China simultaneously building outwards.
Earlier today, we saw Donald Trump and Nvidia chief Jensen Huang in a charming little exchange at the All-In Summit in Los Angeles.
Huang was on stage when the President conveniently rang his phone. He put Trump on speaker and smirked as they waved away the increasingly frantic warnings coming from other industry leaders.
“The robots will not be taking over,” Trump told him. “The AI will not be taking over the rest of the world. The whole thing is a hoax.”
Just over a week ago, Huang declared “AGI has arrived” following the release of OpenAI’s GPT-6 Astra, spruiking his own company’s crucial chips. So crucial, that they have turned Nvidia into a behemoth of unprecedented scale.
It was outside the global top 100 a decade ago and worth just $US58 billion ($A81 billion) by the end of 2016. The AI boom catapulted it from roughly $US560 billion ($A785 billion) five years ago to more than $US5 trillion ($A7 trillion) and the world’s No. 1 spot today.
So, according to one of the biggest financial beneficiaries of the AI boom, machines have already crossed one of computing’s most dangerous and consequential thresholds — but questioning where that leads is unscientific nonsense.
Huang later adopted a slightly more careful position, saying whistleblowers should be taken seriously and companies should slow down if they believed their systems were moving beyond control. But he rejected extinction scenarios as “not grounded in science”.
It’s quite clear that nobody — not Huang, not Trump, not Altman and not the frazzled safety researchers — can say with confidence what these systems will be doing in a decade.
The International AI Safety Report 2026 says researchers cannot reliably predict when specific capabilities will arrive, while expert opinion on the risk of losing control ranges from “implausible” to potentially catastrophic.
Yet the public is being offered two swaggering absolutes by some of the planet’s most powerful voices … either the machines will inevitably lead to our extinction, or only an idiot would worry.
But then there’s Burry, who is telling everyone to simply follow the money.
‘There is nothing to slow down’
Burry was immortalised in The Big Short after profiting from the collapse of America’s subprime mortgage market. His remarkable story was made into a hit movie, and now hordes of fans take his advice as gospel.
As the panic starts to set in, Burry believes the industry’s sudden plea for restraint deserves a much more cynical reading.
“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.
His case has four parts.
Number one, he outright rejects the premise that the current large language models on the market will eventually become artificial general intelligence.
“LLMs are not AI and won’t be AGI,” he wrote on X. “There is nothing AI to slow down.”
Large language models are overwhelmingly classified as artificial intelligence by researchers, companies and governments. But what remains fiercely disputed is whether scaling them will produce AGI, a system able to outperform humans across most economically valuable intellectual work.
Second, Burry argues that slowing development would protect the incumbents that have already spent billions establishing an enormous lead.
“Competition is coming up fast, slowing benefits incumbents,” he wrote.
What he is describing is regulatory capture, where a company spends a fortune reaching the front of the race, then supports rules expensive and complicated enough to stop smaller rivals catching it.
Burry also believes apocalypse warnings could be doubling as advertising.
“IPOs need hype & puffery; ‘we are so awesome it could become dangerous’ is hype & puffery,” he wrote.
He then suggests the safety push provides “cover for real uncontrollable slowing growth as IPOs look to be pushed out”.
Altman has ruled out an OpenAI float in 2026, telling Fortune that listing now would be an “ill-advised moment” because the company needs freedom to prioritise safety over shareholder returns.
Anthropic, meanwhile, is still expected to proceed with a public listing this year, according to Axios. It could argue that a listing provides more transparency.
Cynics will argue that a deliberate slowdown would also slash its bills while restricting new competition, and Burry is far from alone in questioning the motives.
White House technology adviser David Sacks told AI companies to “stop pretending the motivation to slow down is purely altruistic”.
Venture capitalist Bill Gurley said that if the industry were regulated, its leaders should not be allowed to choose the regulators or write the rules.
Vice President JD Vance called companies begging Washington to regulate them a potential “trojan horse”. Trump went further, branding the safety campaign a “SICK conspiracy”.
Huang has also suggested the surge in cyber warnings may partly be connected to companies preparing to sell new security products.
But Burry’s argument rests on the proposal that being an extremely useful technology does not automatically justify an extremely high valuation, asking why they suddenly want the rulebook written after building such an enormous lead.
The trillion-dollar question
The eye-watering fortune involved in the AI game makes healthy suspicion mandatory, given the risks.
Stanford University’s 2026 AI Index estimates global corporate AI investment reached US$581.7 billion (A$815 billion) in 2025 alone, more than double the previous year.
Private AI investment hit $344.7 billion (A$483 billion), while generative AI companies attracted US$170.9 billion (A$240 billion).
The International Energy Agency says capital spending by only five major technology companies exceeded US$400 billion (A$561 billion) in 2025 and is expected to rise another 75 per cent this year — implying a bill of roughly US$700 billion, or almost A$1 trillion.
Five companies are now spending more than the entire world invests in oil and gas production.
Data-centre electricity consumption is projected to almost double from 485 terawatt-hours in 2025 to 950 terawatt-hours by 2030. AI-focused centres are growing even faster.
The cost of frontier development is also becoming truly absurd.
The International AI Safety Report estimates a leading training run already requires roughly $500 million (A$700 million) of computing resources. The next generation could cost between US$1 billion and US$10 billion — about A$1.4 billion to A$14 billion each.
The report openly warns that simply throwing more resources at models may produce diminishing economic returns.
A recent Bank for International Settlements working paper went even further.
Its model estimated the AI race had driven investment about 50 per cent above the socially efficient level under its conservative assumptions, potentially rising to three times the efficient level if demand proves less responsive than expected.
Thoughts? alexander.blair@news.com.au