Anthropic Mathematician Says AI Proved Riemann Hypothesis After Record IPO

Anthropic held the largest initial public offering in history on Thursday; by Friday one of its own mathematicians said the company’s AI had proven the Riemann hypothesis, the 167-year-old problem that carries a $1 million Clay Mathematics Institute prize.

  • The claim: Levent Alpöge, a mathematician at Anthropic, says he and the Claude model have produced a proof of the Riemann hypothesis, plus the Generalized and Extended versions.
  • The format: The argument exists only in Lean, a proof-assistant language, because a conventional write-up would exceed 250,000 pages.
  • The market signal: Anthropic shares, ticker ANT, fell sharply after the IPO, according to reports.

Alpöge posted the claim on his personal website Friday. He wrote that he guided Claude toward combining two mathematical approaches, after which the system “flew, and danced, and sparkled” past the finish line, outpacing his own ability to follow. The resulting proof is so large he released it solely as Lean code, which a computer can mechanically verify for logical consistency.

Lean verification is not the same as human peer review. James Maynard, a mathematician at the University of Oxford, told the Times he is “98 or 99 percent certain” the work is correct, but stressed the computer verification itself must be carefully examined.

Hector Pasten of the Pontifical Catholic University of Chile told Scientific American a confirmed proof would complete a “dictionary between different branches of math where the connections were only partial.”

Anthropic priced its offering Thursday under the ticker ANT. The IPO was the largest in history, yet the stock “precipitously declined” in its debut session.

The Clay Mathematics Institute has not commented on the submission. Its rules require publication in a refereed journal and general acceptance by the mathematical community before the $1 million prize is awarded. Alpöge acknowledged the work is not yet independently confirmed.

Anthropic Mathematician Says AI Proved Riemann Hypothesis After Record IPO
Photo: Yahoo

Tom’s Hardware said that earlier this year OpenAI helped create a breakthrough solution for the Navier-Stokes problem, a finding somewhat overshadowed by a plagiarism controversy. That episode has made mathematicians cautious about AI-generated “proofs” that arrive before human experts can absorb them.

Until a human referee signs off, the Riemann hypothesis remains officially unsolved, and Anthropic’s post-IPO stock decline remains unexplained by the claimed breakthrough.