OpenAI Just Claimed a Huge Math Discovery. Some Academics Are Crying Foul

Details

OpenAI announced that a new AI system solved the Navier-Stokes partial differential equation, one of six remaining Clay Millennium Prize problems each carrying a $1 million award. The solution was formalized in Lean, a programming language for machine-verifiable mathematical proofs. OpenAI researcher Sebastien Bubeck said the company began training a new math-focused AI model on August 28, then mobilized over 1,000 agents working for more than 50 hours after hearing rumors that Anthropic was nearing a solution. Mark Chen, head of research at OpenAI, confirmed compute costs ran 'in the millions of dollars.' The announcement was immediately contested. NYU mathematician Tristan Buckmaster, along with researcher Levent Alpöge, alleges that OpenAI became aware of their in-progress solution and raced to publish first—then attempted to influence who would receive formal credit for the work. This raises substantive questions about ethics, academic norms, and how AI labs should interact with researchers working in the same problem space. If independently verified, this would be the first AI-proven Clay Millennium Problem and arguably the most significant mathematical breakthrough attributable to AI systems. The controversy over credit attribution, the role of competitive intelligence between AI labs (OpenAI vs. Anthropic), and the multi-million-dollar compute investment required make this a pivotal story regardless of how the credit dispute resolves.