A wider route for research resources

Anthropic says it has expanded its AI for Science program through partnerships with 20 foundations, institutions and research groups. The October 28 update moves beyond support aimed primarily at individual scientists.

The announcement describes three routes: funders and societies distribute Claude credits, institutions use credits within their organizations, and specific research projects receive them directly. The company names the Simons Foundation, EMBL and Princeton’s FlyWire among the participants.

Credits provide access to a company’s service. They are not the same as unrestricted research cash, and an announced allocation does not show how much useful scientific work it will produce.

The credit program is not new

Anthropic launched AI for Science on May 5, 2025. That original announcement offered API credits to selected institution-affiliated researchers and described review by company staff, including subject specialists.

API credits pay for usage through a software interface. They should not be confused with an automatic Claude subscription for every researcher at a named institution.

The new institutional network changes distribution and reach. It is also distinct from the short guest-investigator collaborations Model Current covered on October 24, which proposed working directly with Anthropic researchers.

What a named partner confirms

The Simons Foundation published its own announcement on October 15, before Anthropic’s latest update. It describes a partnership with Anthropic and OpenAI committing $15 million in combined funding and resources.

Simons says it will manage and award funding, including support for workshops and initiatives helping mathematicians evaluate AI-assisted results and develop shared standards. It also says the companies will supply tools and computing resources to selected Simons-supported researchers, including in basic science.

That is primary confirmation from one participating foundation. The combined figure is not $15 million in cash newly awarded by Anthropic on October 28, and it does not establish the size of the wider credit expansion.

Allocation and impact are separate questions

Anthropic’s expansion notice provides no total credit value or uniform allowance per institution. It says many collaborations are still at an early stage.

Researchers still need the relevant organization’s allocation rules rather than assume a place on the partner list guarantees individual access. Model Current has not reviewed every institution’s agreement or independently confirmed all 20 relationships.

The useful next evidence would show who received resources, what they used, which tools or datasets were released and whether scientific findings survived normal validation. A broader support network creates opportunities for that work; it is not itself a research result.

Sources

  1. Anthropic: AI for Science institutional expansionOctober28 primary company update:20 named organisations/groups,three distribution routes and early-stage caveat. Size/scope and future impact remain attributed; no total credit value or common per-institution allocation stated. No blanket individual entitlement inferred.
  2. Anthropic: original AI for Science programMay5,2025 primary launch:free API credits,qualified institution-affiliated researchers,company review with subject experts. Historical context; not a newly founded programme in October2026.
  3. Simons Foundation: mathematics and AI partnershipOctober15,2026 primary partner announcement independently confirms Simons/Anthropic/OpenAI partnership,$15m combined funding and resources,foundation-managed awards and selective tools/computing support. Not new cash-only Anthropic funding on October28 or confirmation of every other listed partner.