Google DeepMind supports U.S. Department of Energy on Genesis: a national mission to accelerate innovation and scientific discovery

Google DeepMind just signed on as a partner in the Genesis Mission, a White House-backed effort to rewire how American science gets done using AI. The pitch, laid out by Pushmeet Kohli and Tom Lue, is that the Department of Energy's 17 National Laboratories will get direct, accelerated access to DeepMind's research tools, starting today with something called AI co-scientist.

AI co-scientist is a multi-agent system built on Gemini that's designed to act less like a chatbot and more like a research partner — combing through literature, proposing hypotheses, and drafting research plans. DeepMind says it's already flagged drug repurposing candidates for liver fibrosis that held up in lab testing, and predicted antimicrobial resistance mechanisms that later matched published results before those papers even came out. That's the kind of claim that sounds impressive on a blog post; whether it holds up across thousands of DOE researchers with messier, less curated problems is the real test.

The rollout doesn't stop there. Early 2026 brings AlphaEvolve, a coding agent for designing algorithms that DeepMind credits with squeezing extra efficiency out of its own data centers and chip layouts. AlphaGenome, aimed at decoding the non-coding regions of DNA, gets folded in too, with DeepMind floating future uses in crop resistance and biofuels. WeatherNext, already tied into National Hurricane Center forecasting, rounds out the initial toolkit, alongside general access to Gemini for Government and Gemini 3 for labs that want a broader assistant rather than a specialized model.

DeepMind leans hard on its AlphaFold history to justify the trust here — pointing out that Brookhaven National Laboratory's Protein Data Bank work underpinned AlphaFold, which eventually won Demis Hassabis and John Jumper a share of the 2024 Nobel Prize in Chemistry. It's a fair point: national lab data has quietly fueled some of the biggest AI wins in science already. The new deal is DeepMind betting that pattern repeats at scale, with fusion energy, materials discovery, and earth science named as the next targets for joint research.