NIH Asked Whether to Cap PIs at 2, 3, or 4 Grants. It Got 3,300 Answers — and the Most Common One Was: Cap Dollars, Not Grants.
August 28, 2026 · 6 min read
Granted Research Team · Editorial policy
NIH asked a yes-or-no question with three numeric options. The research community answered with a different question entirely, roughly 3,300 times.
In August 2026, NIH's Office of Extramural Research published its summary of responses to NOT-OD-26-086, the Request for Information issued June 8, 2026 proposing to cap the number of simultaneous Research Project Grants a single principal investigator may hold. The comment period closed August 3. Nearly 3,300 responses came in — an extraordinary volume for an RFI, and a signal in itself.
The most-proposed alternative was not "yes at three" or "no cap." It was cap total dollars rather than grant count.
What NIH proposed
The RFI laid out limits of 2, 3, or 4 simultaneous RPGs per PI, with NIH's own modeling of what each threshold would redistribute:
| Cap level | Funding freed | Additional investigators supported |
|---|---|---|
| 4 simultaneous RPGs | ~$1.3 billion (about 4.8% of RPG funding) | ~1,900 |
| 3 simultaneous RPGs | ~$2.0 billion | ~3,020 |
| 2 simultaneous RPGs | ~$3.5 billion | ~5,200 |
In FY2025, over 10 percent of NIH PIs held three or more simultaneous RPGs. That concentration is the premise of the entire proposal.
NIH offered two implementation paths. The first is a phased reduction, in which over-cap investigators come into compliance gradually through competing renewals, PI role changes, and restrictions on accepting new awards. The second is a hard deadline, requiring full compliance within roughly one year via relinquishment or transfer of PI status.
The stated rationale had three parts: broaden geographic and institutional distribution of NIH funding, increase support for early-stage and mid-career investigators, and improve oversight and mentorship quality by reducing the number of projects any one PI is nominally responsible for.
What the community actually said
The ranked alternatives in NIH's summary are the substance of the comment record:
- Cap dollars rather than grant count — roughly 750 responses
- Partially count multiple-PI grants — roughly 700 responses
- Enforce existing overlap review policies and processes — roughly 700 responses
- Exempt small grants such as R03 or R21 — roughly 670 responses
- Require minimum PI effort per grant — roughly 400 responses
- Cap total funding per lab — roughly 250 responses
Read as a group, these are not scattered objections. They converge on a single technical critique: grant count is a poor proxy for the thing NIH says it wants to manage.
A PI holding four R21s at $275,000 in total direct costs each controls less money and less institutional footprint than a PI holding two large R01s with substantial supplements. Counting awards treats those identically. Counting dollars does not. Requiring a minimum effort commitment per award — say 10 or 15 percent — targets the actual concern about oversight capacity more precisely than counting the number of lines on a biosketch.
The overlap-review comment is the most pointed. NIH already has authority, through Just-in-Time other support review and scientific/budgetary overlap policy, to decline to fund an award where a PI's committed effort or scientific scope duplicates existing support. Roughly 700 respondents effectively told NIH it does not need a new rule; it needs to use the one it has.
The three cross-cutting concerns
NIH's summary flagged three issues raised by both supporters and opponents of a cap. Each is a real behavioral prediction, and each is worth taking seriously.
Team science and multi-PI collaboration would contract. If every multiple-PI award counts fully against every named PI's cap, the rational move for a senior investigator near the ceiling is to decline MPI roles. That directly penalizes the collaborative structures NIH spent fifteen years building, and it falls hardest on the junior partner in an MPI pairing — who is frequently the early-career investigator the policy is meant to help. The 700 responses proposing partial counting of MPI awards are aimed exactly here.
Formal PI roles would shift while actual control stayed put. A capped investigator can move a grant's contact-PI designation to a senior postdoc, a research assistant professor, or a junior faculty member while continuing to direct the science, control the budget, and occupy the lab space. This is the most predictable evasion pathway, it is nearly impossible to police from grant records, and it creates a class of nominal PIs whose independence is a paper fiction — which is worse for their careers than not holding the award at all.
Applications would get bigger. If you can hold fewer awards, each award should be larger. That is straightforward optimization. The likely outcome is consolidation of what would have been three modular R01s into one large program-project-scale request, which reduces the number of funded investigators — the opposite of the policy's goal — while concentrating risk in fewer, larger reviews.
What NIH signaled about next steps
NIH characterized the outcomes as high-level and said the comments will inform next steps. No cap level has been selected, no implementation path chosen, and no effective date announced.
That is meaningful. NIH ran an RFI, received a comment record heavily weighted toward alternatives rather than toward its own three options, and published a summary that foregrounds those alternatives rather than dismissing them. Agencies that intend to proceed unchanged do not usually structure a summary that way.
The plausible outcomes, in rough order of likelihood based on the record:
- A dollar-denominated threshold rather than a count — a total-direct-costs ceiling per PI, likely with a review trigger rather than a hard bar.
- A count-based cap at four with exemptions for small mechanisms and partial counting for MPI awards — the least disruptive version of the original proposal, redistributing the smallest amount.
- Strengthened overlap enforcement dressed as a policy notice rather than a new cap, which requires no new authority.
- No action, with the proposal absorbed into the broader unified funding strategy discussion NIH has been running through 2026.
A hard two-RPG cap on a one-year compliance deadline is the least likely outcome and would be the most disruptive by a wide margin.
What to do now
Do not restructure your portfolio in anticipation. There is no rule. Relinquishing an award or transferring PI status to get under a hypothetical cap is an irreversible decision made against a policy that may never issue, and may issue in dollar terms that make your grant count irrelevant.
Do audit your committed effort honestly. Whatever form this takes, minimum-effort requirements and overlap enforcement are the common denominator across every likely outcome. A PI listing 5 percent effort across eight awards is exposed under a count cap, a dollar cap, a minimum-effort rule, and existing overlap policy simultaneously. Bringing committed effort into line with actual effort is the one preparatory step that is correct regardless of which version lands.
If you are early-career, do not plan around the redistribution. Even the most aggressive scenario frees money on a multi-year phase-in, and the freed dollars flow through ordinary competitive review — not through a set-aside. The 1,900 to 5,200 additional investigators are a modeling output, not a reserved pool. Your R01 strategy should not change.
If you are a research office, model the MPI exposure at your institution now. Query your NIH portfolio for investigators at three or more RPGs and, separately, for investigators who appear as non-contact PI on multiple MPI awards. Those two populations are the ones whose behavior changes first under any version of this policy, and the second group is invisible in most standard portfolio reports.
Watch for the policy notice, not the news coverage. The operative document will be an NIH Guide notice with an effective date tied to application due dates or budget periods. Until that exists, everything else — including this analysis — is inference from a comment summary.
The larger frame
The grant cap proposal did not arrive in isolation. It sits alongside the six-application-per-year submission cap we covered in NIH Just Capped Grant Applications at Six Per Year, the proposed changes to reporting peer review outcomes under NOT-OD-26-088, and a broader agency effort to manage a constrained budget by rationing access rather than by growing the pot.
Each of these individually is arguable. Together they describe an agency that has concluded it cannot fund the research enterprise at its current shape and is redistributing scarcity. The 3,300 responses to NOT-OD-26-086 are, in the aggregate, the community saying: if you are going to redistribute scarcity, at least measure it correctly.