$293 Million to Merge AI With National Science: Inside the DOE Genesis Mission

Nearly $300 million, 26 challenge areas, and a nine-page acronym of ambition — the Department of Energy's Genesis Mission is the largest single AI-for-science funding call the federal government has ever issued. And the first deadlines are five weeks away.

Announced on March 17, Under Secretary for Science Darío Gil framed the initiative as something more than a typical grant solicitation. "The Genesis Mission has caught the imagination of our scientific and engineering communities to tackle national challenges in the age of AI," he said. The vision is to assemble interdisciplinary teams — spanning national laboratories, universities, industry, and philanthropic science organizations — that use novel AI models to solve problems ranging from quantum algorithm discovery to nuclear reactor optimization.

For researchers and small companies looking to break into DOE's orbit, this is an unusually accessible entry point. But the two-phase structure, the breadth of topics, and the compressed timeline demand a clear-eyed understanding of what DOE is actually buying.

The Shape of $293 Million

The Genesis Mission operates on a two-phase model that borrows the staged investment logic familiar to anyone who has navigated SBIR or DARPA programs.

Phase I awards range from $500,000 to $750,000 for nine-month projects. These are small-team efforts — proof-of-concept demonstrations that an AI-driven approach can credibly address one of the 26 designated challenge areas. Think of Phase I as a feasibility sprint: can your team show that a novel AI model or framework produces results that traditional methods cannot?

Phase II jumps dramatically — $6 million to $15 million over three years, with mandatory partnerships between national laboratories and either industry or academic collaborators. These are the large-team, infrastructure-grade investments that DOE expects to drive operational breakthroughs.

Crucially, applicants can enter at either phase. Teams with mature AI capabilities and existing DOE relationships may skip Phase I entirely and submit a Phase II letter of intent. Teams new to the DOE ecosystem can use Phase I to build the track record and partnerships that make a Phase II bid viable.

Total program funding stands at $293.76 million, with cost sharing required — a detail that will shape how industry participants structure their proposals.

The 26 Challenges: What DOE Is Actually Looking For

The Genesis Mission is organized around nine broad topic areas, subdivided into 26 specific challenges. The topics span:

• Advanced manufacturing — using AI to accelerate materials discovery and optimize fabrication processes
• Biotechnology — AI-driven protein engineering, synthetic biology, and bioenergy research
• Critical materials — reducing dependence on foreign supply chains for rare earth elements and strategic minerals
• Nuclear fission — AI for reactor design optimization, safety analysis, and fuel cycle modeling
• Nuclear fusion — plasma control, materials degradation prediction, and pathway-to-commercialization modeling
• Quantum information science — two challenge areas that deserve special attention (see below)
• Semiconductors and microelectronics — AI-assisted chip design and manufacturing process optimization
• Discovery science — fundamental physics, chemistry, and materials science accelerated by AI
• Energy — grid optimization, renewable integration, and carbon management

The quantum challenge areas stand out for their specificity. Challenge 7 ("Discovering Quantum Algorithms with AI") asks teams to use AI workflows to find efficient error correction codes, develop fault-tolerant algorithmic tools, and explore quantum advantage in domains like plasma physics, nuclear systems, and lattice quantum chromodynamics. Challenge 8 ("Realizing Quantum Systems for Discovery") targets hardware — AI-enhanced device design, real-time calibration systems, multi-qubit sensing, and decoherence mitigation for scalable quantum networks.

For teams working at the intersection of AI and quantum computing, these two challenges represent the federal government's clearest statement yet that the path to quantum advantage runs through machine learning.

Who Can Apply — and Who Should

The eligibility requirements are broad: any U.S. institution can apply, including universities, national labs, private companies, and science philanthropies. But the competitive dynamics favor different applicants at each phase.

Phase I is designed for agility. A university research group with a novel AI architecture and domain expertise in, say, critical materials discovery could submit a compelling $500K proposal without needing a national lab partner. The nine-month timeline rewards teams that can move fast and produce demonstrable results.

Phase II tilts toward established consortia. The $6–15 million award range and three-year timeline signal that DOE expects Phase II teams to include national laboratory scientists who can provide access to DOE computing infrastructure, experimental facilities, and classified datasets. Industry partners who can contribute cost share and commercialization pathways will strengthen any Phase II bid.

Small businesses and startups should pay particular attention. Unlike many DOE Office of Science programs, the Genesis Mission does not restrict eligibility to large institutions. A small AI company with a differentiated model architecture and a partnership letter from a national lab could compete effectively in Phase I — and use that award to build the credibility needed for Phase II or for separate DOE SBIR proposals.

Timeline and How to Position

The clock is ticking:

• March 26, 2026 — Informational webinar (attend this; DOE webinars routinely surface unstated evaluation priorities)
• April 28, 2026 — Phase I applications and Phase II letters of intent due
• May 19, 2026 — Full Phase II applications due
• December 17, 2026 — Final application closing date

For Phase I, the five-week window between now and April 28 means teams need to be assembling partnerships and drafting technical approaches now. The most competitive proposals will demonstrate three things: a genuinely novel AI approach (not incremental improvements to existing models), deep domain expertise in the specific challenge area, and a clear path from Phase I results to Phase II scale.

For Phase II, the letter of intent due April 28 is the gating mechanism. DOE will use LOIs to screen proposals before inviting full submissions in May. The LOI should establish the team's credentials, the scope of the proposed work, and — critically — the national lab partnerships that anchor the project.

What Makes Genesis Different

The Genesis Mission is not DOE's first foray into AI for science. The Office of Science has funded AI research through its Advanced Scientific Computing Research program for years, and DOE announced $68 million in separate AI-for-science funding in 2025. The agency's national laboratories — Argonne, Oak Ridge, Lawrence Berkeley, Sandia, and others — collectively operate some of the most powerful computing infrastructure in the world.

What makes Genesis different is the explicit mandate to organize AI investment around specific outcome-oriented challenges rather than open-ended research. DOE is not asking teams to "explore AI applications in energy" — it is asking them to solve named problems within defined timeframes using AI as the primary tool.

The appointment of Darío Gil as both Under Secretary for Science and Genesis Mission Director reinforces this operational focus. Gil, who spent 32 years at IBM including a decade leading IBM Research, brings a private-sector mindset about translating research into deployable technology. His involvement signals that DOE will evaluate Genesis proposals not just on scientific merit but on their potential to produce results that can be transitioned to operational use within DOE's mission space.

Strategic Considerations for Applicants

Pick your challenge narrowly. With 26 challenge areas, the temptation is to propose something that spans multiple domains. Resist it. DOE reviewers will be organized by challenge area, and the most competitive proposals will demonstrate deep expertise in a single challenge rather than broad competence across several.

Lead with the AI architecture. The Genesis Mission is fundamentally about whether novel AI approaches can accelerate scientific discovery. Proposals that treat AI as a supporting tool for traditional research methods will be less competitive than proposals where the AI methodology is the central innovation.

Understand the partnership expectations. Phase I does not formally require multi-institutional teams, but the RFA's emphasis on interdisciplinary collaboration means that solo-institution proposals need to explain why a single team can cover the necessary ground. Phase II explicitly requires national lab involvement.

Attend the March 26 webinar. DOE informational sessions routinely reveal unstated priorities, clarify ambiguous language in the NOFO, and provide signals about how proposals will be evaluated. Missing the webinar is a competitive disadvantage.

Budget for cost sharing. The Grants.gov listing confirms cost sharing is required. Industry applicants should plan for this from the outset; university applicants should check whether their institution's research office can provide matching commitments.

The Genesis Mission represents the federal government's most ambitious bet that AI can transform how science is done — not someday, but within the next three years. For researchers and companies working at the frontier of AI and physical science, the April 28 deadline is the one to circle. Tools like Granted can help teams identify which challenge areas align with their capabilities and build proposals that match DOE's expectations before the window closes.