The Simons Foundation announced on 15 October 2026 that, under a partnership with Anthropic and OpenAI, the three will commit $15 million in combined funding and resources to help mathematicians navigate AI. The foundation will manage and award the money, including through an open call for proposals.
This is a community-facing grant and workshop programme, not a new model release. OpenAI said in a 15 October update on its mathematics page that it is contributing $5 million to the Simons initiative. Anthropic’s exact dollar share is not stated on the Simons release or OpenAI’s update.
What the money is for
According to the Simons Foundation, awards will support workshops and initiatives that help researchers:
- evaluate recent AI-assisted results
- learn and use new tools
- share perspectives
- shape collective standards for AI use in mathematics
OpenAI’s update adds that funding managed by Simons will support workshops, conferences and special programs, and that OpenAI will also give Simons-supported researchers access to its models and computing resources.
Simons says Anthropic and OpenAI will likewise provide AI tools and computing resources to researchers the foundation supports, extending beyond mathematics into basic science. Programme deadlines and eligibility details are not in the press note; the open call is still to come.
Spergel: the community should drive adoption
Simons Foundation president David Spergel said the moment is for the mathematics community to come together. In a quote on the foundation’s release: “The community should be driving discussions of how these tools are evaluated, adopted and used, not any single company or institution.” He also framed the effort as looking at mathematics “as a collective understanding driven by human curiosity,” not only as a pile of proofs.
That community-led line is the political heart of the announcement: frontier labs are co-funders and tool providers, while Simons positions itself as the awarder and convener so standards are not written solely inside one company.
Why Simons is the intermediary
The foundation stresses its long backing of mathematics and of machine learning for science, including early support for the Lean proof assistant — the formalisation stack that makes machine-checkable, AI-assisted proofs practical. It says it has already convened workshops and collaborations on AI’s role in the field. Putting an open call under Simons is meant to route money and compute through a research philanthropy rather than a vendor grant programme alone.
What this is not
OpenAI’s mathematics page still hosts its 6 October 2026 release of AI-produced math and computer-science results (the dump TSN covered separately). This post is only about the 15 October funding partnership. It does not re-litigate those results, Lean formalisation counts or independent verification status.
Anthropic has not published a matching dollar figure on the Simons page. Treat the $15 million as the combined commitment of funding and resources; treat OpenAI’s $5 million as the only disclosed cash split so far.
Context from earlier TSN coverage
AI-for-maths has moved quickly from demos to contested research artefacts. TSN covered OpenAI’s release of hundreds of math and CS results, where Lean helped and independent checks were still pending; DeepMind’s AlphaProof Nexus work in Science; and a preprint crediting Shanghai AI Lab’s Intern-S2 with help on Ghys’s conjecture. The Simons programme sits downstream of that wave: less “here is a new theorem dump,” more “how should the field evaluate and standardise AI-assisted work?”
Why it matters
Frontier labs want mathematicians to use their tools; mathematicians want evaluation norms that are not vendor-owned. A $15 million pot managed by Simons, with model and compute access attached, is a concrete attempt to fund that middle layer — workshops, standards and scrutiny — while OpenAI and Anthropic still show up as resource providers. For basic science beyond pure maths, the same access clause could matter if Simons routes support outside mathematics departments.
What’s still unknown
- Anthropic’s cash share of the $15 million combined commitment
- Open-call timeline, eligibility and award sizes
- How compute/model access will be gated and for how long
- Whether standards emerging from the workshops will bind lab disclosure practices
What to watch
Watch the Simons open call when it opens, and any separate Anthropic note that clarifies its contribution. Also watch whether workshop outputs cite evaluation standards that labs then adopt for future result drops.
What this means
- $15 million combined from Simons, Anthropic and OpenAI, managed by Simons, with an open call still to come.
- OpenAI: $5 million plus model/compute access for Simons-supported researchers.
- Anthropic’s dollar split not stated; tools and compute access are promised alongside OpenAI’s.
- Spergel’s line: the community, not any single company, should drive evaluation and adoption.
Sources
- Simons Foundation, “Simons Foundation Launches $15M Effort To Help Mathematicians Navigate AI Era”, 15 October 2026
- OpenAI, “Sharing AI progress in mathematics” (Update, 15 October 2026)
Related stories
- OpenAI’s 372 math and CS results: Lean helps, independent checks pending
- Google DeepMind’s AlphaProof Nexus Is Published in Science: 9 of 353 Erdős Problems, 44 of 492
- A Preprint Credits Shanghai AI Lab’s Intern-S2 With Helping Prove Ghys’s Codimension-One Conjecture
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
