Simons Foundation launches $15M effort to help mathematicians navigate AI era

The announcement came on 15 October 2026, and the partners are Anthropic and OpenAI. The foundation will run an open call for proposals and fund workshops and initiatives on evaluating recent AI-assisted results, learning new tools, and setting collective standards for AI use. The release gives only the combined total and no split by partner. OpenAI said separately that it is contributing $5 million. Anthropic and OpenAI will also supply AI tools and computing resources to foundation-backed researchers, with uses reaching beyond mathematics into basic science. Foundation president David Spergel said the aim is for the mathematics community, not any single company or institution, to decide how AI tools are assessed and adopted. The foundation earlier funded development of the Lean proof assistant, which made AI-assisted proofs possible. One reading of the program is that it targets checking AI output more than producing it. For AI companies selling into research, verification is the bottleneck this money targets. Formal tooling like Lean and community-set standards are likely to shape which math claims get accepted. Labs are also using subsidized compute and model access to build ties with academic users, so products that make outputs checkable may find an audience here.