In July 2026 the White House Office of Science and Technology Policy released Science: A New Golden Age, a manifesto that openly models itself on Vannevar Bush's 1945 report and proposes to tilt hundreds of billions in federal research emphasis away from universities and toward industry, artificial intelligence, and the individual scientist. It matters because this is the intellectual scaffolding for how the administration wants to restructure a roughly $200 billion federal R&D enterprise, with direct consequences for NIH, NSF, and biomedical research. The caveat, raised by supporters and critics alike in STAT's coverage, is the distance between what the report preaches and what the same administration is actually doing.

What it is. A diagnosis paired with a prescription, structured across five chapters and an annex of FY2028 R&D budget priorities. It sets four goals: back the individual scientist over the legacy institution, change how grant dollars are allocated and assessed, couple discovery to domestic manufacturing, and re-engineer the research enterprise for the AI era. Think white paper, not experiment.

What it argues, with the numbers it leans on:

• Federal R&D runs about $200 billion a year. Private industry now deploys roughly $700 billion, more than triple government and higher education combined, and its share of basic research is rising.
• Drug development has gotten steadily less efficient. The report cites "Eroom's law," the finding that pharmaceutical R&D productivity in new drugs per billion dollars has fallen roughly eighty-fold since 1950, halving about every nine years, even as the NIH budget more than doubled since the 1990s.
• Researchers spend nearly half of their federally funded time on paperwork. Indirect cost rates at NIH-funded institutions average north of 40 percent, while those same institutions often accept 10 to 15 percent from private funders.
• China has reached rough parity with the United States on R&D spending, adjusted for purchasing power.
• Around half of U.S. doctorates in computer science and mathematics now go to temporary visa holders.

What it proposes. Fund people rather than projects (expand NSF's Graduate Research Fellowship, scale NIH Pioneer-style awards). Diversify how money goes out the door: "golden tickets" for reviewers to champion long-shot ideas, fast grants, prize challenges, ARPA-style program managers, new X-Labs. Restore "permissionless innovation." Launch the Genesis Mission, a flagship AI-for-science initiative. And enforce reproducibility through a "Restoring Gold Standard Science" executive order.

The problems it names are real, and scientists across the political spectrum have been complaining about most of them for years. Administrative burden is genuine. Eroom's law is genuine. Review panels really do reward consensus and punish the unconventional proposal. Funding people rather than projects is not a MAGA invention; it is the Howard Hughes Medical Institute model, and NIH already runs versions of it. The document is well written, historically grounded, and in places refreshingly candid about the reproducibility crisis. Stripped of its politics, a fair amount of it would draw nods at any faculty lunch.

Reform-minded voices agree. Alec Stapp of the Institute for Progress, writing at Marginal Revolution, singled out five ideas he considers overdue: metascience units to let agencies experiment on their own funding, a portfolio approach in place of one-size-fits-all project grants, independent X-Labs for ambitious challenges, fast grants and prize mechanisms, and a serious push to give researchers their time back. These are real proposals with real constituencies. The disagreement is rarely about the diagnosis.

The life sciences quietly leave the room. For a report this concerned with drug development, the biomedical enterprise is treated mostly as a cautionary tale about inefficiency. The named priorities are physical sciences, AI, robotics, nuclear, semiconductors, and quantum. The advisory board has zero biomedical members. If you run a dialysis company or take care of kidney patients, the "golden age" on offer is not obviously yours. Jeremy Berg, the former director of NIH's National Institute of General Medical Sciences, told Forbes the report is "tremendously misinformed," applying a venture-capital template to federal science that federal science was never meant to fit.

The diagnosis and the actions point in opposite directions. The report praises the NSF Graduate Research Fellowship while the administration's own budgets propose cutting it in half. It preaches gold-standard science while, per the Union of Concerned Scientists, 194 federal advisory bodies have been terminated. You cannot credibly promise more rigor while dismantling the machinery that produces it.

The core inference is a leap. Eroom's law is real, but "therefore shift emphasis from universities to industry" does not follow. The report's own figures concede that industry's $700 billion is overwhelmingly late-stage development, which is precisely the work markets already fund. The basic research that markets will not touch is the part universities and the NIH exist to do. Bush's original point, quoted approvingly in the report, cuts against the report's own conclusion. As Vanderbilt astrophysicist Keivan Stassun put it to Forbes: "You don't get more apple pie by cutting down the apple trees."

Some of the appealing phrases carry freight. "Permissionless innovation" sounds liberating until you remember it is being applied to pharmaceuticals. "Fund the person" reduces accountability at the same time it reduces bureaucracy. And several serious observers read the whole document as the conceptual foundation for an OMB rule injecting political judgment into what is supposed to be merit review. That worry is not abstract. At a congressional hearing the same week, Kratsios insisted the White House "is not involved in which proposals get funding" at NSF, yet House Democrats displayed an agency workflow chart reading "Coordination with DOGE and WH take place outside of workflow," tied to roughly $1 billion in redirected NSF funds (FedScoop). Meanwhile the report bets heavily on AI at a moment when 59 percent of adults say they do not trust AI companies to act responsibly.

Take it seriously as a diagnosis and with a large grain of salt as a plan: the problems are real and some of the fixes are sensible, but the gap between the rhetoric and the administration's simultaneous cuts, plus the near-total absence of a biomedical voice, means this golden age may not have much room for the life sciences.