White House launches the Genesis Mission for AI science

Back in the 1940s, the Manhattan Project not only produced the atomic bomb, but also laid the groundwork for the modern Department of Energy and its sprawling national laboratory system. Now, the White House believes the global race for AI dominance demands an effort of the exact same scale.

The White House issued Executive Order 14363, officially launching the Genesis Mission. The concept is straightforward: mobilize decades worth of federal scientific datasets, pair them with the nation's most powerful supercomputers, and build specialized AI foundation models and agents to tackle America's hardest scientific problems.

An all-in platform for government AI

At the center of this initiative is a new infrastructure project called the American Science and Security Platform. Led by the Department of Energy with overarching guidance from the Assistant to the President for Science and Technology, the platform is designed to unify federally funded research power.

First, it pulls together supercomputing infrastructure across national laboratories along with secure cloud environments. Second, it unifies massive, federally curated datasets that have taken decades and billions of taxpayer dollars to compile. Finally, it aims to develop domain-specific AI foundation models and autonomous research agents capable of testing hypotheses, running workflows, and even operating robotic laboratories for autonomous physical experimentation.

Washington is not moving slowly here. Within 60 days, the Secretary of Energy must submit a list of at least 20 national science and technology challenges. The order specifically singles out advanced manufacturing, biotechnology, critical materials, nuclear fission and fusion, quantum information science, and semiconductors.

From there, the clock keeps ticking. Federal computational resources must be inventoried within 90 days, followed by initial data assets at 120 days. By day 240, officials must review robotic lab capabilities across facilities, leading to an initial operating demonstration on at least one national challenge within 270 days.

Why it matters

We think this executive order signals a major shift in how the US government views AI. It is no longer just treating machine learning as Silicon Valley commercial software, but as the core engine of national strategic power and scientific sovereignty.

For scientists, researchers, and enterprise partners, this could eventually unlock access to goldmines of unique scientific data that private industry could never recreate. Imagine training chemistry or materials models on decades of classified or curated lab experiments. If executed well, this could massively shorten the research cycle for clean energy and advanced chips. That said, bureaucracy and strict national security classifications mean early access will likely be heavily gatekept.

Placing the Department of Energy at the wheel is a clever move. Unlike purely regulatory agencies, the DOE already controls the world's fastest supercomputers and actual physical research plants. However, a major question mark remains around funding. The order repeatedly notes actions are subject to available appropriations. Without fresh, dedicated congressional funding, turning an ambitious AI Manhattan Project into reality may prove much harder than signing the order.