Inside the DOE Genesis Mission: How a \$293M FOA Is Reshaping How the Department of Energy Funds AI for Science

The Department of Energy ran one of the most consequential federal AI funding solicitations of 2026 with comparatively little public attention. The Genesis Mission RFA, posted as DE-FOA-0003612 on March 17, 2026, closed its Phase II application window at 11:59 p.m. Eastern on May 19, 2026 — three days before this analysis published. With $293,760,000 in total available funding, award ceilings reaching $16 million per project, and a topical scope spanning 21 research areas and 99 focus areas, Genesis Mission is the largest single AI-for-science procurement the federal government has run this fiscal year. It is also, in important respects, a structural innovation that signals where DOE's Office of Science is heading over the next decade.

For the principal investigators, national lab teams, and university research consortia that submitted Phase II applications on May 19, the immediate work is done. For everyone else — and particularly for teams that won Phase I awards in late April and are now staring at a December 17, 2026 Phase II from Phase I deadline — the strategic work is just beginning. This deep dive is for both audiences: a careful reading of what the Genesis Mission actually is, what its design choices reveal about DOE's funding philosophy in 2026, and how teams positioned for the next round should be thinking about the months ahead.

The Architecture of the Solicitation

Genesis Mission is structured as a two-phase competition with parallel and sequential tracks. The architecture is unusual and worth understanding in detail because it differs meaningfully from the standard DOE Office of Science FOA model.

Phase I (small team applications) opened in March and closed April 28, 2026. Phase I awards target compact research teams pursuing a quantitative assessment of whether a proposed AI approach is on a credible trajectory to deliver transformative scientific capability or transformative impact on an energy application. Phase I funding sits at the lower end of the $500,000 to $16 million award range, with individual awards typically expected in the $500,000 to $2 million band. The deliverable is not a complete scientific result. It is a defensible early-stage demonstration that the approach is worth scaling.

Phase II (large team applications) ran on a parallel track. Phase II Letters of Intent were due April 28, 2026. Full Phase II applications closed May 19, 2026. Phase II awards target larger consortia — typically anchored by a national laboratory with one or more university partners — pursuing more mature scientific or technical capability development. Phase II awards can reach the $16 million ceiling and typically fund three to five years of work.

The Phase II from Phase I track is the design innovation that deserves the closest reading. Phase I awardees can apply for Phase II awards by a December 17, 2026 deadline, meaning teams that win Phase I funding in summer 2026 will have approximately five to six months to convert their early demonstrations into a credible Phase II proposal. This mirrors the SBIR Phase I-to-Phase II pipeline, but at orders of magnitude greater scale.

Across all tracks, the FOA solicits work in 21 topics and 99 focus areas spanning advanced manufacturing, biotechnology, critical materials, nuclear fission, nuclear fusion, quantum information science, semiconductors and microelectronics, discovery science, and the broader energy portfolio.

What the Topic Selection Reveals

The 21-topic, 99-focus-area structure is itself a strategic signal. DOE's Office of Science is not pursuing AI as a horizontal capability to be applied wherever it might fit. The Genesis Mission topic list is a deliberate, vertical-by-vertical articulation of where DOE program officers think AI can credibly deliver transformative outcomes on a five-year horizon.

Three observations stand out.

Nuclear fission and nuclear fusion are separate topics. This is the first major DOE solicitation to treat them as distinct AI research portfolios rather than bundling them under a generic nuclear-energy umbrella. The split implies that DOE expects materially different AI architectures and datasets for fission optimization, advanced reactor materials, and fuel-cycle modeling than for the inertial confinement and magnetic confinement plasma physics work that dominates the fusion portfolio. Research teams pursuing nuclear AI work need to choose which side of that line their work falls on and tailor their proposals accordingly.

Quantum information science is treated as a science topic, not just a computing capability. The Genesis Mission framing treats quantum systems as a domain where AI can accelerate discovery — particularly in error correction, noise modeling, and the design of new qubit architectures — rather than treating quantum as an alternative computing substrate competing with AI. This framing aligns with NSF's recent $1.5 billion X-Labs initiative, which makes similar architectural assumptions.

Critical materials and semiconductors get parallel attention. The inclusion of both critical materials and semiconductors and microelectronics as distinct topics reflects the CHIPS and Science Act priorities flowing through DOE's research portfolio. Teams with expertise in materials discovery for advanced semiconductors are uniquely positioned because they can credibly propose into both topic categories with thematically related but distinct work programs.

The Award Range — Why $500K to $16M Matters

The 32-to-1 ratio between the minimum and maximum award sizes is unusually wide for a DOE Office of Science solicitation. Most DOE FOAs run a much tighter range — perhaps 4-to-1 — which produces a more uniform award portfolio. The Genesis Mission's wide range reflects an explicit decision to fund both compact, focused exploratory work and large multi-institution consortia within a single solicitation.

This has two practical implications for applicants.

Award size signals expected portfolio role. A team applying for a $500,000 to $1 million Phase I award is implicitly proposing to do early-stage capability demonstration. A team applying for a $10 to $16 million Phase II award is proposing to operate as one of the anchor institutions in DOE's AI-for-science portfolio for the next several years. The proposals look fundamentally different. A team that proposes mid-range work at the $5 to $8 million Phase II level needs to articulate why that scale is right for the scientific question — not too small to demonstrate transformative impact, not so large that it crowds out the agility advantage of a focused team.

Cost share and matching expectations vary across the range. Larger Phase II awards anchored at national labs come with built-in DOE infrastructure that effectively absorbs much of the indirect cost burden. Smaller Phase I awards going to university PIs need to budget realistically for indirect costs, which can consume 25% to 40% of the total award depending on the institution. Teams should be reading their target award level alongside their indirect cost rate to understand the actual research dollars available.

The Hidden Story: Cooperative Agreements vs. Grants

Genesis Mission is funded through a mix of cooperative agreements and project agreements rather than traditional research grants. The distinction is technical but consequential.

A cooperative agreement implies substantial DOE program officer involvement in the conduct of the research — not just funding oversight, but technical direction, milestone setting, and ongoing scientific engagement. For PIs accustomed to NSF-style grants where the program officer is largely hands-off after award, the cooperative agreement structure of Genesis Mission requires a different working relationship. DOE program officers will be deeply involved in shaping the trajectory of the research, including potentially redirecting work as scientific results emerge.

The project agreement variant — used for some Phase II awards — provides somewhat more PI autonomy but still implies closer DOE engagement than a traditional grant. Teams considering applying to Genesis Mission in future rounds should plan for the operational reality of close DOE involvement and design their work plans accordingly.

What Phase I Awardees Should Be Doing Right Now

For teams that submitted Phase I applications by the April 28 deadline and are awaiting award decisions in early summer 2026, the next five months will determine whether they convert Phase I funding into a Phase II proposal that wins. The December 17, 2026 deadline sounds far away. It is not. Realistic Phase II proposal preparation requires four to six months of concerted effort, which means Phase I awardees need to be working on Phase II planning effectively from the moment they receive their Phase I award letter.

Three priorities deserve immediate attention.

Demonstrate AI advantage, not just AI application. Phase II proposals will be evaluated against the standard articulated in the FOA: a "quantitative assessment of whether the proposed approach to demonstrating AI advantage in the scientific and/or technical research workflow is on a trajectory to yield a transformative scientific capability or transformative approach to an energy application." That phrasing — "AI advantage" — is doing significant work. Teams that demonstrate that they used AI to do something scientifically useful will lose to teams that demonstrate AI delivered an outcome that would have been infeasible by conventional approaches. The Phase I window is the moment to design experiments that produce that quantitative AI-advantage signal.

Build the consortium structure for Phase II. Phase II awards favor larger consortia anchored by a national laboratory with university and industry partners. Phase I awardees coming from university-only teams need to identify and engage national lab partners during the Phase I performance period. This is not a partnership that can be assembled in the final month before the December 17 deadline. National labs operate on long planning cycles, and the staff time required for a credible Phase II application needs to be negotiated and committed by late summer at the latest.

Maintain rigorous documentation. Genesis Mission Phase II proposals will require detailed documentation of Phase I results, including quantitative performance metrics, validation against benchmark approaches, and clear articulation of what the proposed Phase II work would build on top of the Phase I findings. Teams that treat Phase I as exploratory work without disciplined documentation will struggle to assemble the evidence base Phase II reviewers expect.

How Genesis Mission Fits the Broader Federal AI Funding Landscape

Genesis Mission is one of three major federal AI funding shifts that have materialized in spring 2026. The NSF X-Labs program brings $1.5 billion in ten-year funding with innovative Other Transaction Authority mechanisms. The NSF-NVIDIA $152 million open AI for science partnership emphasizes fully open models and training data. Genesis Mission, at $293 million, sits in the middle — larger than the NSF-NVIDIA partnership but smaller than X-Labs, and substantially more domain-specific than either.

The cumulative picture is a federal AI-for-science portfolio that has substantially expanded in 2026 even as some other federal research budgets have contracted. Teams positioning for the next round of any of these programs should read the three together. The topical priorities — quantum, materials, advanced manufacturing, biology — are remarkably consistent across the three. The funding architectures differ, but the underlying scientific bets are aligned.

For Granted readers tracking these shifts, the operational implication is straightforward. The federal AI-for-science funding environment in 2026 is unusually accessible to teams that can credibly demonstrate quantitative AI advantage in priority domains. Genesis Mission's Phase II from Phase I deadline on December 17, 2026 is the next major milestone. Phase I awardees should treat the intervening months as the most important strategic window in the program.

The full FOA document remains available on the DOE Office of Science FOA page. Teams considering applications in future Genesis Mission cycles — and DOE has signaled additional rounds are likely — should be reading the current solicitation carefully and beginning the consortium-assembly work now.