5,000 applications. That’s 2.5 times the previous record for a DOE funding call, and it’s the number of proposals that landed in Darío Gil’s inbox after the first Genesis mission announcement in March. Next month, the agency will pick a limited set of winners and hand out $293 million in collective funding.
Genesis: The AI Platform That Connects 17 National Labs
The $600 million Genesis mission, launched last November by executive order, tasks the DOE with building a single AI platform that can query scientific instruments, supercomputers, and datasets across all 17 U.S. national labs. The goal is to give researchers—from universities and private companies alike—a unified architecture to tackle hard scientific problems. Gil says the platform is already demonstrating multi-stage autonomy: AI agents accessing government data, launching supercomputer simulations, and controlling a manufacturing robot to 3D-print a part. Humans still in the loop, but the pipeline is real.
Fault-Tolerant Quantum by 2028: 26 Problems That Matter
Last week the DOE announced its push for the world’s first fault-tolerant quantum computer solving scientific problems by 2028. That’s a direct response to Trump’s quantum innovation executive order. The agency has defined 26 science and technology challenges of national importance that the Genesis platform must address by August 21. One of them: “re-envision” how AI agents orchestrate end-to-end research. Gil claims the DOE is on track to meet that deadline.
Skepticism and Funding Fights: “I Haven’t Moved One Dollar”
Basic-science researchers worry that Genesis money is coming out of nuclear and high-energy physics budgets. Gil pushes back: the total DOE Office of Science budget per research area is up year-over-year, and the pinch scientists feel comes from rising costs of maintaining user facilities, not from AI. The Genesis contribution across two rounds is $520 million, with some cash drawn from basic-science programs. Gil’s line: “All I am asking is that if you’re doing biology, can you please explore carefully and thoughtfully the implications of this computing revolution for biology?”
On AI safety, Gil won’t touch regulation directly—he points to the Trump administration’s hands-off approach—but notes that deep scientific collaboration at the frontier of AI is a better mechanism than just reacting. That’s a veiled reference to the recent forced takedown of Anthropic’s Fable 5 model.
By August 21, the DOE must show its AI platform solving at least one of those 26 national challenges. Gil says they’re on track. For the rest of academia, the message is clear: if you want DOE funding now, your proposal better have an AI angle.
Source: Trump has big AI and quantum ambitions: this scientist's job is to make them reality
Domain: nature.com
Comments load interactively on the live page.