Anthropic puts $150M behind federal science labs

Scientific breakthroughs usually require two things in massive quantities: funding and compute. Anthropic is stepping in to cover a big chunk of the latter.

At the White House Office of Science and Technology Policy's summit, Anthropic announced a major commitment: $150 million over three years to support the federal Genesis Mission, an initiative aimed at accelerating scientific discoveries with artificial intelligence.

Claude heads to federal labs

Under the deal, Anthropic is deploying Claude, Claude Code, and heavy API credits across more than 15 federal agencies, including NASA, the National Institutes of Health (NIH), and the National Science Foundation (NSF).

This isn't starting from scratch. Anthropic first teamed up with the Department of Energy last December to bring Claude into national laboratories. This new expansion spreads the tools to several hundred research projects across the government.

Anthropic says the effort will zero in on some of the toughest technical problems on the table, with fusion energy and quantum computing sitting right at the top of the priority list.

Beyond handing over API keys, the company will also provide hands-on onboarding, training, and technical support to help agencies launch and maintain their first AI-assisted research workflows.

A deliberate push into the sciences

The Genesis Mission commitment fits into a much wider campaign Anthropic has been quietly executing all year.

Earlier this year, the company introduced Claude Science, a dedicated AI workbench built around the common tools and software packages that researchers actually use, alongside 10,000 free or discounted seats for academic scientists.

They also kicked off a research preview for the Model Hardware Standard. That project aims to build a universal safety specification allowing AI agents to directly control physical lab instruments.

Why it matters

For government scientists, this clears out the bureaucratic red tape and budget constraints that usually make accessing frontier models a headache. Getting direct access to Claude Code in projects like fusion plasma modeling or quantum simulation offers immediate practical value without dipping into agency grant money.

For Anthropic, this is a very smart enterprise play. While consumer chatbots fight an expensive war for attention, Anthropic is entrenching Claude deep inside foundational American scientific infrastructure. Giving away $150 million in credits and compute across three years is relatively cheap compared to the long-term enterprise lock-in it creates across 15 high-profile agencies.

The flip side to watch is dependency. When national labs build their daily experimentation and data pipelines around proprietary models, switching later becomes painful and expensive. We think Anthropic is playing the long game here, aiming to ensure that the next decade of scientific discoveries runs on its rails.