In National Institutes of Health v. American Public Health Association (No. 25A103, 21 August 2025), a 5–4 Supreme Court stayed a district court order that had restored roughly $783 million in cancelled research grants — not by ruling the cancellations lawful, but by holding that money claims against the government belong in the Court of Federal Claims, not district court. The finding of illegality was left standing and stripped of effect. Read as a mechanism rather than a culture-war episode, the case shows how forum selection now does the work a ruling on the merits used to do.
On 21 August 2025 the Supreme Court did something quieter, and more durable, than deciding who was right. In National Institutes of Health v. American Public Health Association (No. 25A103), a 5–4 majority stayed a district court order that had reinstated roughly $783 million in cancelled NIH research grants — not by ruling the cancellations lawful, but by ruling that the wrong courtroom had heard the case (Supreme Court order; Crowell & Moring; Science/AAAS). Nearly a year on, the shape of that move is fully visible, and it is the more important story. Established
Two months earlier, in June 2025, District Judge William G. Young had found the underlying policy unlawful. NIH's directives cancelling grants tied to diversity, gender identity, and COVID-19 research violated the Administrative Procedure Act; Young went further, calling the terminations "racial discrimination, and discrimination against America's LGBTQ community" (Higher Ed Dive; Science/AAAS). He vacated the guidance and ordered the money restored — some 2,500 terminations undone. That was a substantive administrative-law defeat for the executive. Established
The Move That Mattered
The Supreme Court did not disturb Young's finding that the guidance was likely unlawful. It left that vacatur standing. What it stayed was the part with teeth: the order restoring the grants. Its reasoning was jurisdictional. Under the Tucker Act, claims for money against the federal government belong exclusively in the Court of Federal Claims, not the district courts. A suit to force NIH to pay out cancelled grants, the majority held, is at bottom a money claim — whatever constitutional label the plaintiffs attach to it. As Justice Barrett put it, "the APA's waiver of sovereign immunity does not extend to enforcing contractual obligations to pay money, regardless of the substantive basis for the claim" (Supreme Court order; Crowell & Moring). Established
Strip away the subject matter and the instrument is clean. Jurisdiction-channelling does three things at once, and none of them is a ruling on whether the executive acted lawfully.
First, it removes the remedy that bites. The Court of Federal Claims can award money after the fact; it generally cannot order an agency to reinstate a grant or reverse a policy. The one thing grantees actually need — the funding turned back on — is precisely what the forum they are sent to cannot compel (Vinson & Elkins; State Impact Center). Established
Second, it converts a question of substance into a question of routing. Young answered "did the executive break the law?" The Supreme Court answered "which court may ask?" — and let the first answer stand untouched while draining it of effect. The merits were not overturned. They were made irrelevant to the outcome.
Third, it changes the clock. A district court can vacate a policy for everyone at once. The Court of Federal Claims processes money claims one plaintiff at a time, slowly, after the loss has already been absorbed. Litigants must now run what one firm called a "two-step dance" — an APA suit in district court for the policy, a separate Tucker Act suit in the Court of Federal Claims for the money — while 28 U.S.C. § 1500 bars them from pursuing both on the same facts at once (Vinson & Elkins). Established
The net effect: the executive won the operational question — the cuts stood — while the legal question it had lost stayed lost on paper and off the table in practice. Justice Jackson, in dissent, called the framework a "labyrinth" that preserves "the mirage of judicial review" (Supreme Court order; Vinson & Elkins). That is a precise description of the mechanism, not merely a complaint about the result. Assessed
The Five Who Chose the Forum
The alignment is worth reading, because it was not the familiar one. The majority to stay the grant order comprised Thomas, Alito, Gorsuch, Kavanaugh, and Barrett; Chief Justice Roberts joined Sotomayor, Kagan, and Jackson in dissent on that point (Supreme Court order; Crowell & Moring). Barrett supplied the controlling jurisdictional rationale. Gorsuch and Kavanaugh, concurring in part and dissenting in part, would have gone further still. The disagreement among the nine was not really about the research topics. It was about how far the Tucker Act reaches — about which door a plaintiff must use — and that argument is portable to any grant, contract, or subsidy the government chooses to cancel. Established
A Year Later, the Grind
From the vantage of August 2026, the second-order effects are no longer hypothetical. In December 2025 NIH settled with a coalition of states and researchers. On its face a climbdown, the terms show the mechanism still holding: NIH agreed to re-review frozen and denied applications under ordinary scientific process rather than the struck-down directives — but the deal "did not require that NIH fund any specific application," and set decision deadlines running into mid-2026 (STAT; Higher Ed Dive). Re-review is not restoration. The agency conceded the process and kept the outcome. Established
This is the part that should concern institutions far from this case. American research universities and teaching hospitals built operating budgets on the assumption that federal grants, once awarded, were durable — and that if an administration cancelled them unlawfully, a court could turn them back on quickly. NIH v. APHA unbundles that assumption. The policy may still be struck down; the money still routes through a forum that pays slowly, individually, and only in dollars, never in reinstatement. An institution can be legally right and operationally broke at the same time. Assessed
What to Watch
The question is not whether grantees eventually recover money in the Court of Federal Claims; some will. It is whether the two-step structure becomes the standard executive answer to any adverse grant ruling — cancel, lose on the merits, and win the year it takes to route the remedy elsewhere. Watch how quickly the Court of Federal Claims moves on the first NIH money claims, and whether other agencies begin citing APHA when their own terminations are challenged. The instrument was built here. Whether it generalizes is the institutional story of the next term. Assessed