$15 million to keep up with the machines
On October 15, 2026, the Simons Foundation announced a new set of initiatives to help the mathematics community navigate the rapid rise of AI in the field. The Simons Foundation, Anthropic, and OpenAI will commit $15 million in combined funding and resources. The foundation will manage and award the funding — including through an open call for proposals — to support workshops and initiatives that help researchers evaluate recent AI-assisted results, learn and use new tools, share perspectives, and shape collective standards for AI use. Anthropic and OpenAI will also provide AI tools and computing resources to researchers supported by the foundation, extending beyond mathematics into basic science.
Why now: the 722-manuscript flood
The timing is no coincidence. Just days earlier, on October 6, an internal OpenAI frontier model tackled roughly 4,000 problems and released 722 manuscripts organized into 372 families of results — each averaging about three hours of intensive compute on ChatGPT Pro. Only 162 carried full Lean formalizations that allow computer verification of the main claim. The rest remain provisional, some already required corrections, and three were withdrawn shortly after release. Mathematicians were rattled: the field went from scarcity of results to a deluge of unverified ones overnight.
“As artificial intelligence rapidly reshapes the research landscape, we want to help the mathematical community come together and forge the best path forward. Now is the time to look at mathematics as a whole — not as a collection of proofs and theorems, but as a collective understanding driven by human curiosity.”
What's in the $15M package
- Open call for proposals, managed and awarded by the Simons Foundation
- Workshops and conferences bringing researchers face-to-face with AI-generated results
- Targeted programs to evaluate recent AI-assisted results
- Anthropic: funding plus computing resources and Claude access for supported researchers
- OpenAI: funding (reportedly $5M) plus model access and substantial compute
- Community-driven awards and shared standards for how AI tools are evaluated and used
The bigger picture: AI is changing what counts as research
This is the same week CERN and Mistral announced a Large Physics Model for particle physics. Two $15M-scale science-AI partnerships in one day is a pattern: AI labs are buying legitimacy and data access in basic science, and science institutions are buying influence over how the tools get used. For creators, the honest angle is the verification gap — the space between "the model produced it" and "we believe it".
How to prompt AI research-visualization images for this story
The visual story is verification, not magic: mathematicians checking machine output, formal proofs under review, the flood of manuscripts versus careful human scrutiny. Prompt images about the process, not the hype.
Editorial illustration for an AI-and-mathematics story: a vast dark library hall where floating pages of mathematical proofs cascade down like a waterfall from above; at a long wooden table below, mathematicians review printed pages with red verification marks while one glowing screen shows a formal Lean proof being checked line by line; warm lamplight against the dark hall, cinematic, thoughtful, no text rendered.
Frequently Asked Questions
Who is funding the $15M initiative?
The Simons Foundation, Anthropic, and OpenAI jointly commit $15 million in combined funding and resources. OpenAI contributed $5 million plus model access and compute; Anthropic made a similar commitment including Claude access. The Simons Foundation manages and awards the funding.
Why do mathematicians need this now?
On October 6, 2026, an OpenAI frontier model released 722 manuscripts at once — only 162 with full Lean formal verification. The field was flooded with provisional results, some needing corrections, and needed infrastructure to absorb, test, and build on them.
What is Lean formalization?
Lean is a proof assistant — software that lets mathematicians write proofs in a formal language a computer can check mechanically. Only 162 of the 722 AI-generated manuscripts carried full Lean formalizations, meaning the rest can't be machine-verified as-is. The Simons Foundation funded Lean's early development.
How can researchers apply for funding?
The Simons Foundation will manage the funding through an open call for proposals, supporting workshops and initiatives. Awards are community-driven — watch the Simons Foundation's announcement channels for the open call details.
Does this extend beyond mathematics?
Yes. The announcement says Anthropic and OpenAI will provide AI tools and computing resources to Simons-supported researchers extending into basic science — and it landed the same day as the CERN-Mistral Large Physics Model partnership.
