Claude Claims Historic Math Breakthrough After Anthropic IPO

Mathematician Levent Alpöge says Claude helped crack a 167-year-old conjecture, though independent verification is still underway. Because the announcement dropped right after Anthropic’s IPO, observers are questioning the unusual timing of the release.
A researcher at Anthropic reports that Claude helped solve the Riemann hypothesis, a 167-year-old problem to which a $1 million reward is attached.
The announcement arrived Friday, just a day after the AI firm’s Wall Street debut, though the findings have not yet been verified by outside experts.
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,”
Levent Alpöge wrote in a four-page note.
He notes that this technique also solves the wider Generalized Riemann Hypothesis, through which the concept is extended to additional mathematical functions.
Alpöge states that he and Claude formalized their argument using Lean, a program designed to verify mathematical proofs step by step. While his note provides links to 14 downloadable sections of the proof archive, public availability alone does not guarantee acceptance by the wider math community.
Alpöge notes that he contacted Anthropic CEO Dario Amodei regarding a potential announcement “late Sunday night” on November 15. After consulting with company leadership, he mentioned that a specific constraint had to be respected, though he stopped short of explaining what it actually was.
“The exact timing is awful,” he admitted. While Alpöge pushed back against cynical readings of the schedule, whether the delay was driven by the company’s IPO was never explicitly confirmed by him.
Shares of Anthropic started trading on the Nasdaq on Thursday, November 19, after $81.34 billion was raised during its initial offering.
The “Mathocalypse”
Alpöge also played a central role in September’s AI math rivalries. OpenAI announced it had solved the Navier-Stokes Millennium Prize problem while he and NYU mathematician Tristan Buckmaster pursued related fluid-equation research, triggering a dispute in which credit over research priority was contested.
OpenAI followed up in October with 722 manuscripts produced by an unreleased model, which were organized into 372 families of results. Just 162 featured Lean-verified main findings, prompting mathematicians to demand stronger proof rather than taking the company’s wave of claims at face value.
Some mathematicians worry about an impending “mathocalypse,” where coveted problems are solved by AI before researchers can learn from tackling them. Toronto mathematician Daniel Litt cautioned that, without adaptation, “there’s just no more math in 50 years.” Speaking at an MIT panel this week, Buckmaster maintained that mathematics still relies on human insight.
Why a Verified Mathematical Proof Would Matter
Initially proposed by Bernhard Riemann in 1859, the conjecture forecasts where specific zeros of the Riemann zeta function lie. The problem carries weight because those zeros are tied to the distribution of prime numbers, whose patterns form the foundation for vast areas of mathematical study.
Confirmed proof would mark a landmark achievement in mathematics, given that the problem has been pursued by researchers for more than a century.
Lean can confirm that formal deductions follow its rules. However, the archive still must be reproduced by mathematicians, who also need to verify that the formalized theorem aligns with the original conjecture and scrutinize the claimed result before treating it as settled.
Anthropic had previously shared advancements regarding the zeta function without claiming to resolve the entire hypothesis. Alpöge’s latest assertion reaches well beyond what was accomplished in that earlier work.
The Clay Mathematics Institute includes the Riemann hypothesis among its Millennium Prize Problems, each offering a $1 million reward. According to the organization’s guidelines, a proposed solution has to appear in a qualifying publication, endure at least two years of review, and achieve broad consensus before the prize is considered.
This article follows our editorial policy. Spotted an error? Request a correction. Not financial advice.


