The White House chose a familiar frame for its most ambitious science bet in a generation: the Manhattan Project. When officials gathered at a July 22 summit to announce the first winners of the Genesis Mission, the administration's flagship artificial-intelligence initiative, they billed the program as an effort "comparable in urgency and ambition" to the wartime crash effort that built the atomic bomb. This time the target is not a weapon but the pace of American discovery itself, and the tool is AI trained on the government's vast, siloed troves of scientific and health data.
Roughly 5,000 teams applied. About 280 projects were selected in this first phase, spanning all 50 states, and each will receive a nine-month grant of between $500,000 and $750,000 to point AI at concrete problems, from stabilizing the US power grid to automating laboratories. Winning teams can also compete for a second phase: a three-year program carrying millions of dollars in annual funding. The administration has said it is committing more than $5 billion to the effort overall.
The rollout capped a consequential stretch for the administration's science agenda. A day before helping to unveil the grants, President Trump's chief science adviser, Michael Kratsios, released a report titled "Science: A New Golden Age" that proposes substantial changes to how the federal government funds research. Together, the two events sketched a strategy: redirect money toward the science the administration favors, above all AI, while rethinking the machinery that has funded American research for decades.
What the Genesis Mission is building
The Genesis Mission was created in November 2025 through Executive Order 14363, which directed the Department of Energy to assemble what the order describes as an integrated platform stitching together the world's fastest supercomputers, national experimental facilities, AI systems and unique federal datasets. The stated goal is to double the productivity and impact of American research within a decade. The Department of Energy anchors the effort through its 17 national laboratories and their machines, and the initiative spans a long roster of federal agencies including HHS, NSF, EPA, the Department of Veterans Affairs and the Department of Defense.
The health applications draw the sharpest picture of the ambition. Under the plan, the Department of Health and Human Services would combine access to the nation's longitudinal health cohorts with the EPA's chemical-monitoring data and DOE's compute to probe the origins of chronic disease. HHS would feed an integrated pediatric-cancer data ecosystem into DOE supercomputers to train models across hundreds of rare childhood-cancer subtypes. And the VA would pair electronic health records and genomic data from its Million Veteran Program with DOE computing to build models that flag disease and health risks earlier. The connective tissue in each case is the same: national data plus national compute, coordinated across agencies that rarely share either.
The winning projects announced in July are more granular. Jeremy Zucker, a computational scientist at Pacific Northwest National Laboratory, is leading a project that will use AI to optimize the growth of algae for products including fertilizers and biofuels. "I was very, very, very happy" to be chosen "because it was incredibly competitive," Zucker told Nature at the summit. Other teams will tackle grid reliability and lab automation.
A strategy, and a fight over its direction
What makes this moment notable is context. For roughly a year and a half, the administration has drawn headlines mainly for trying to cut funding it dislikes, withholding money from universities and reshaping agency budgets. The Genesis grants move in the opposite direction, advancing the science the administration wants to accelerate. Kratsios's accompanying report makes the underlying philosophy explicit, calling for shifting funding away from what he terms "legacy institutions" such as universities. "Too much of our research enterprise has come to serve itself rather than the scientists within it," he wrote, arguing that US science is stagnating and floating alternatives such as "golden tickets" that let individual reviewers override colleagues to back high-risk proposals.
That vision landed unevenly. At a hearing of the House Science Committee the same week, Republicans broadly backed Kratsios's approach while Democrats attacked the administration's budget cuts and its push to exert political control over grant awards. "You aren't launching us into 'the golden age of innovation.' You're sending us back to the Stone Age," said Representative Gabe Amo, a Rhode Island Democrat. Outside government, critics were blunter still; University of Washington biologist Carl Bergstrom wrote that the administration seemed determined to "destroy" US universities.
The oversight questions sharpen around where the money comes from. Nature has reported that the NSF pulled back $300 million from two divisions, rescinding or cutting some 150 proposals that had already passed peer review. Agency staff, speaking anonymously, told Nature the funds appear headed to a separate White House effort called Grand Research Challenges, run out of Kratsios's office. Asked at the hearing about the clawbacks, Kratsios said, "I'm not familiar with those details. I would refer you to NSF," while arguing the country should be "more deliberative and specific" about tackling grand challenges. The NSF declined to comment, saying it "revises budgets and manages portfolios as needed."
That is the tension at the heart of the Genesis Mission. Directing national compute and cross-agency data at pediatric cancer or veterans' health is the kind of coordinated moonshot many researchers have long wanted. But funneling money toward top-down "grand challenges" by pulling back peer-reviewed grants strikes at the bottom-up, investigator-driven model that has defined American science.
What to watch
The near-term test is Phase 2. Which of the roughly 280 first-round teams graduate into multiyear, multimillion-dollar funding will reveal what the administration actually rewards once the pilot money is spent. Watch, too, whether DOE's promised compute, including new AI clusters coming online at the national labs, materializes on schedule, since the whole model depends on it. And watch Congress, which controls appropriations: whether lawmakers ratify the reallocation from core NSF programs, or fight it, will determine whether "A New Golden Age" becomes federal policy or remains a manifesto.
"Too much of our research enterprise has come to serve itself rather than the scientists within it"- Michael Kratsios, White House Chief Science Adviser