Anthropic expands AI for Science program to 20 global institutions

Scaling AI for scientific discovery

Anthropic has expanded its AI for Science program by partnering with 20 global scientific foundations, institutions, and research groups. This initiative provides free Claude credits to researchers who are working on high-impact scientific projects. The company launched the program in May 2025 with the goal of increasing access to AI tools to help scientists deepen their work.

The program aims to accelerate progress across a wide variety of disciplines, including biology, medicine, physics, chemistry, climate science, and mathematics. By providing these resources, Anthropic hopes to enable researchers to tackle some of humanity's toughest problems more efficiently. The company has already supported hundreds of projects through this initiative over the last few months.

Scientists using the tools have reported that they are using Claude to generate new hypotheses and design experiments. They also use the technology to swiftly analyze complex data sets and communicate their final findings. This shift puts scientists at the center of the transformation, allowing them to set the questions and develop the projects that drive the field forward.

Institutional and funding partnerships

A significant portion of the expansion involves partnering with major science funders and societies. These organizations play a key role in finding and funding original research. Anthropic is sharing credits with groups like the Alfred P. Sloan Foundation and the Simons Foundation to distribute resources to ambitious researchers.

The partnership also includes the US National Science Foundation, a federal agency that provides roughly one-quarter of the government's funding for basic research at American universities. Other partners include Schmidt Sciences, which focuses on areas like astrophysics and biosciences, and the Open Source for Science Fund. These groups will help ensure that AI resources reach a broad range of scientific disciplines.

The expansion also reaches into the private and non-profit sectors. For example, the organization is partnering with the Wellcome Sanger Institute, a leading genomics center. They are also working with the Francis Crick Institute in London and the European Molecular Biology Laboratory (EMBL). These institutions will use the credits to enable discovery within their specific research environments.

Specific research initiatives

Anthropic is also providing direct support to several specific research initiatives that focus on problems with large public impact. One such project is FlyWire at Princeton University, which produced the first complete wiring map of an adult fruit fly's brain. This project develops open data and tools for connectomics to help researchers understand how neural circuits create perception.

In the realm of climate science, the LEAP project at Columbia University combines machine learning with physics. This initiative aims to improve how climate models capture complex systems like clouds, oceans, and land. Additionally, the AstroPath project at Johns Hopkins University applies techniques used in space observations to photograph cancer tumors at the individual cell level.

The program also supports genomics and mathematics projects like gnomAD and Tau Ceti. The gnomAD project pools DNA data from over 800,000 people to create an open reference of human genetic variation. Meanwhile, the Tau Ceti project is creating a communal collection of mathematical proofs in Lean, a system that allows computers to verify every step of a proof.

Building public scientific goods

Many of the collaborative projects funded by this program are designed to produce public goods that benefit the wider scientific community. These outputs include open datasets, open-source software, open standards, and shared libraries. By creating these resources, the program seeks to expand what scientists can attempt and lead to new discoveries.

One specific example of this is the development of the Model Hardware Standard (MHS) in partnership with the HHMI Janelia Research Campus. The MHS is a shared specification that allows AI agents to safely operate lab instruments. This standard is currently in a research preview stage and represents a move toward standardized AI interaction with physical hardware.

The Allen Institute is also a key partner in this effort, focusing on producing large-scale public data and tools to help researchers study the brain, cells, and the immune system. The institute is building teams of AI agents to analyze this data. This work aims to turn months of manual analysis into just a few hours of automated processing.

The future of AI in science

As AI continues to change the methodology of scientific inquiry, Anthropic plans to continue supporting and collaborating with scientists. The company intends to expand these efforts to even more partners around the world as the program matures. They are eager to see how these early-stage collaborations can support novel research and scientific discovery.

The goal remains to provide the resources necessary for scientists to deepen and expand their own work. By moving from supporting individual scientists to large-scale institutional partnerships, the program aims to create a more sustainable ecosystem for AI-driven research. This shift helps ensure that the tools are integrated into the foundational structures of science.

The company looks forward to showcasing new ways of working in these scientific fields as the partnerships progress. They hope that by providing these resources, they can help scientists tackle the most pressing challenges facing humanity. The program will continue to evolve as the technology and the needs of the scientific community change over time.