The DOE's AI Pivot Is Starving Traditional Research: What Scientists Should Know About the Genesis Squeeze

A nuclear physicist at a midsize research university opened her email in mid-March and found a funding opportunity that perfectly illustrated her dilemma. The Department of Energy had just posted a $293 million Request for Application under the Genesis Mission — its flagship initiative to merge artificial intelligence with scientific research. The deadline was April 28. She had six weeks to find industry and national laboratory partners, develop a proposal, get it through her university's internal review, and submit. Her existing DOE grant, which funded two graduate students and a postdoc studying neutron-rich isotopes, was expiring in September with no renewal path in sight.

The Genesis Mission did not cause her funding problem. But it accelerated the trajectory that made it inevitable. And her story is repeating across physics departments, chemistry labs, and materials science programs at universities nationwide.

The Numbers Behind the Squeeze

The Department of Energy's Office of Science received $8.4 billion in FY2026 — a substantial budget that makes it the nation's largest funder of physical sciences research. Congress preserved this funding over the administration's objections, rejecting proposed cuts and maintaining the office's position as the primary engine for basic research in physics, chemistry, materials science, and computing.

But how that $8.4 billion is allocated within the Office of Science is shifting. Operations — the cost of running national laboratories, maintaining experimental facilities, and supporting the physical infrastructure of American science — now consume a larger share of the budget than research grants. This imbalance has grown over the past two fiscal years as facility costs have risen and new initiatives like the Genesis Mission have been layered on top of existing commitments.

The impact varies by field, but nuclear physics provides the clearest illustration. The nuclear physics program operates on an $866 million budget. Operations — running facilities like the Continuous Electron Beam Accelerator Facility at Jefferson Lab and the Relativistic Heavy Ion Collider at Brookhaven — account for more than half of that total. Research grants, the money that funds university investigators and their students, have fallen 18 percent. And the Genesis Mission's mandate that agencies redirect a portion of their budgets toward AI-focused research means another slice of the remaining grants budget is being reallocated.

The arithmetic is brutal. When you apply an 18 percent reduction to the grants line and then redirect an additional 10 percent toward Genesis AI initiatives, the money available to renew ordinary grants expiring this year falls short by roughly one-third. One in three nuclear physics investigators whose grants expire in FY2026 will not have funding renewed — not because their science is weak, but because the budget no longer supports the number of active grants it once did.

This is not a nuclear physics problem. It is an Office of Science problem. Energy research, basic energy sciences, biological and environmental research, and high energy physics all face versions of the same arithmetic — rising operations costs and AI mandates compressing the grants budget that sustains university research.

The Genesis Mission's Shadow

The Genesis Mission is, on its own terms, an ambitious and well-designed program. The $293 million RFA invites interdisciplinary teams from national laboratories, industry, and academia to tackle more than 20 national challenges — advanced manufacturing, biotechnology, critical materials, nuclear energy, quantum information science — through novel AI models and frameworks. Phase I awards range from $500,000 to $750,000 for nine-month projects. Phase II awards scale to $6 million to $15 million over three years. The program structure rewards collaboration, and the challenge areas align with genuine national priorities.

Under Secretary for Science Darío Gil has framed Genesis as a transformational moment: "The Genesis Mission has caught the imagination of our scientific and engineering communities to tackle national challenges in the age of AI." The program represents DOE's commitment to positioning artificial intelligence as a force multiplier for scientific discovery.

The problem is not what Genesis funds. It is what Genesis displaces.

When the same Office of Science that funds tens of thousands of individual research grants also launches a $293 million initiative, the money comes from somewhere. Genesis is not funded by new appropriations — it draws from the existing Office of Science budget. Every dollar allocated to Genesis AI challenges is a dollar not available for the investigator-initiated grants that have been the backbone of American physical science research for decades.

The compressed timeline amplifies the displacement effect. DOE announced the $293 million opportunity on March 18 with Phase I applications due April 28 — six weeks for researchers to identify the right challenge area, recruit collaborators at national laboratories and industry partners, develop a technical approach, assemble budgets, and navigate their university's proposal review process. Major university research offices typically require four to six weeks of internal review alone. The timeline functionally limits competitive proposals to researchers who already have Genesis-ready collaborations in place or who can repurpose existing project concepts with an AI overlay.

This creates a two-tier system. Researchers embedded in national laboratory networks — already attending Genesis planning workshops, already collaborating with DOE program managers, already operating at the intersection of AI and their scientific discipline — can pivot quickly. University-based investigators who primarily work in fundamental research, without existing AI collaborations or industry partnerships, face a timeline that makes genuine proposal development nearly impossible.

The AI-Washing Problem

When funding agencies signal that AI is the priority, rational researchers respond by incorporating AI into their proposals — whether or not artificial intelligence is genuinely central to their scientific questions. This phenomenon, which the research community has informally dubbed "AI-washing," is accelerating across DOE programs.

A materials scientist studying corrosion mechanisms adds a machine learning component to predict corrosion patterns. A biophysicist studying protein folding incorporates an AI-driven molecular dynamics simulation. A nuclear physicist analyzing detector data includes an AI-assisted pattern recognition module. In each case, the core scientific question remains unchanged, but the proposal now includes AI vocabulary and budget lines that signal alignment with agency priorities.

AI-washing is not dishonest — machine learning genuinely enhances many research workflows. But when funding pressure forces every proposal to include an AI component regardless of scientific necessity, two things happen. First, reviewers must evaluate AI components that range from transformative to perfunctory, often without the specialized expertise to distinguish between them. Second, the grants budget that should fund the best science in a discipline instead rewards the proposals that most convincingly narrate an AI connection.

The NSF's recent merit review changes — reducing minimum external reviews from three to two and making panel discussions optional — raise parallel concerns about review quality. When the DOE combines AI-priority funding signals with compressed proposal timelines, the risk is that the proposals that succeed are not necessarily the ones with the strongest science, but the ones with the strongest AI narrative and the fastest institutional response time.

Who Wins and Who Loses

The Genesis pivot creates clear winners. National laboratories — already embedded in DOE's research ecosystem, already staffed with computational scientists and AI specialists, already participating in Genesis planning — are positioned to lead Phase I and Phase II teams. Large research universities with dedicated DOE partnership offices, existing AI research centers, and industry collaboration infrastructure can compete effectively. Companies developing AI tools for scientific applications gain a new customer in the federal government.

The losers are equally clear. Mid-tier university research groups that depend on individual investigator grants to fund graduate students and postdocs face a shrinking pool of available funding. Early-career researchers who have not yet built the cross-sector collaborations that Genesis rewards are at a structural disadvantage. And entire subfields of physical science research — areas where AI adds marginal value but fundamental measurement, theory, and experimentation are the core methods — risk becoming unfundable under the current priority structure.

The downstream effects reach beyond principal investigators. Graduate students funded on DOE grants that are not renewed must find alternative support or leave their programs. Postdoctoral researchers face a tighter job market as fewer PI-led projects can afford additional staff. University research infrastructure built around DOE-funded experimental programs — specialized equipment, technical staff, laboratory space — loses its funding justification.

What Researchers Should Do

The DOE's AI pivot is not temporary. The Genesis Mission represents a structural reorientation of the Office of Science's funding priorities, and researchers who treat it as a passing trend will find themselves progressively marginalized.

Build genuine AI collaborations now. The most competitive Genesis proposals will come from teams that have real computational expertise, not bolt-on AI components. If your research could genuinely benefit from machine learning, computer vision, or foundation models, invest in a collaboration with a computational scientist before the next funding cycle. Attend DOE informational webinars (the next Genesis webinar is available through DOE's Office of Science website), visit national laboratory partners, and develop joint proposals with AI researchers who understand your domain.

Diversify your funding portfolio. The Office of Science's grants budget is structurally compressed, and no single appropriations cycle will fix it. Researchers who rely exclusively on DOE grants are increasingly vulnerable. NSF, NIH (for biomedically adjacent work), ARPA-E (for energy applications), and DOD programs all fund physical sciences research with different priority structures. The FY2026 budget scorecard shows that NSF received $8.75 billion and NIH received $47.2 billion — both programs where physical science researchers can compete for funding that is not subject to the same AI mandate.

Maintain your fundamental research program. The pressure to AI-wash is real, but researchers who abandon their core scientific questions to chase AI funding risk losing the expertise that makes them competitive in any program. The most sustainable strategy is to maintain an active fundamental research agenda while developing a parallel track of genuinely AI-integrated work. This requires more proposals, more collaborations, and more administrative overhead — but it preserves the scientific identity that defines your lab.

Engage DOE program managers directly. The Genesis Mission's compressed timelines reward researchers who are already in conversation with DOE about upcoming opportunities. Program managers at the Office of Science can provide guidance on challenge areas, collaboration expectations, and proposal mechanics that are not available in the published solicitation. This is standard practice for DOE grants, but it matters more when the timeline between announcement and deadline is measured in weeks rather than months.

The Structural Question

The DOE's pivot raises a question that extends beyond any single funding cycle: what is the right balance between directed mission programs and investigator-initiated research?

The American research enterprise was built on a model where federal agencies fund the best science proposed by the best scientists, with broad latitude for researchers to define their own questions. Directed programs like Genesis represent a different philosophy — the agency defines the problem, sets the methods (AI), and selects teams to execute. Both models produce valuable science. But when directed programs grow at the expense of investigator-initiated grants, the pipeline of fundamental discoveries that feeds future directed programs begins to thin.

The researchers whose grants expire this year without renewal were not studying frivolous questions. They were measuring nuclear structure, characterizing novel materials, probing the behavior of matter under extreme conditions — the kind of fundamental science that, a decade from now, will be the foundation for AI applications that do not yet exist. When the DOE tells those researchers that their funding has been redirected to AI, it is making a bet that today's AI priorities are worth more than tomorrow's fundamental discoveries.

That bet may pay off. The Genesis Mission could produce breakthroughs that accelerate scientific progress across every discipline it touches. But if it does not — if the AI-integrated projects produce incremental improvements rather than transformative discoveries — the fundamental research programs that were defunded to make room for Genesis will not be easily rebuilt. Graduate students will have moved on. Equipment will have been decommissioned. Expertise will have scattered.

The DOE has $8.4 billion and a mandate to advance American science. The question is whether the current allocation serves that mandate — or whether the AI pivot, however well-intentioned, is consuming the scientific foundation it was meant to enhance.