Michael Burry — known for betting against US housing before 2008 — has argued that names such as Nvidia and Palantir sit near the centre of an artificial-intelligence valuation bubble that could end uglier than the early-2000s telecom wipe-out. In public remarks he described buying two-year put options on both, stressing that headline “notional” exposure reported in the press was far larger than the cash he actually put at risk (on the order of about ten million dollars per position in one account he discussed).
His critique is less about whether AI software is useful and more about market plumbing: capital expenditure announced by hyperscalers that, he says, inflates equity values several times over; GPU depreciation schedules stretched to five or six years while hardware ages out faster; and, on Palantir, profits that look thinner once stock-based compensation is treated as a real cost. Call it a sceptical research note, not a prophecy.
Markets can stay expensive longer than a famous short-seller stays comfortable. Use the warning to check leverage and narrative stocks — including crypto vehicles that ride the same risk appetite — not as a reason to fund a Telegram desk that claims to “trade Burry’s thesis” for a fee.
If a pitch like this cost you money — or you are unsure what to do next — start at Get help.
This article is for general information only. It is not legal or financial advice. Details change. If you have lost money, contact your bank or card issuer and use the routes on the help pages.