Two of the largest private research funders in the world have, within about a year of each other, rewritten what “publishing” means for the scientists they fund. The Howard Hughes Medical Institute’s Immediate Access to Research policy takes effect for articles submitted on or after January 1, 2026. The Gates Foundation’s refreshed 2025 Open Access Policy has applied to funded manuscripts since January 1, 2025. Both are widely shorthanded as “preprint-before-submission mandates” — but that phrase overstates what either policy actually requires, and the difference matters for anyone advising authors on compliance. This page sets out what each funder actually requires, where the plain-English framing and the policy text diverge, and what changed in practice for authors and the research-administration staff who support them.
CASRAI has separate, more detailed coverage of the Gates Foundation Open Access Policy in full; this page focuses on the preprint-timing requirement specifically and on how it compares to HHMI’s newer policy.
What HHMI’s “Immediate Access to Research” policy requires
HHMI issued the policy on September 24, 2025, as a successor to its 2022 Open Access to Publications policy (which remains available as a transition option for articles submitted to a journal before January 1, 2026 but not yet published). It applies to HHMI Investigators, Freeman Hrabowski Scholars, Janelia Group Leaders and Senior Group Leaders, Janelia Scientific Operation Managers, Theory Fellows, Principal Scientists, Janelia Fellows, AI Scientists, AI@HHMI Machine Learning Researchers, and Project Scientists — but only when the article is a “major contribution” from their lab, defined specifically as one where they or a lab member is listed as (co-)first, (co-)last, or (co-)corresponding author.
- Two required preprints, not one. Covered scientists must deposit both an initial preprint (the first formal communication of the findings) and, later, a revised preprint that incorporates peer-review feedback and any new results, before the article is published by a journal.
- Timing is a recommend/require split, not a flat “before submission” rule. HHMI “strongly recommends” depositing the initial preprint by the time of journal submission, but the hard requirement is looser: the preprint must be deposited “promptly after a journal has sent the manuscript to peer reviewers.” In practice this means an author can technically submit to a journal first and still comply, as long as the preprint goes up once peer review begins — the policy text does not make preprinting a precondition of submission itself.
- Designated servers only. bioRxiv is the preferred server; medRxiv (clinical medicine), ChemRxiv (chemistry), and arXiv (computer science and physics) are permitted for out-of-scope work. Other servers are not compliant.
- CC BY 4.0 only. Preprints must carry a Creative Commons Attribution 4.0 International license. HHMI’s policy explicitly lists CC BY-NC, CC BY-NC-SA, CC BY-ND, and CC BY-NC-ND as non-compliant.
- Scope limited to original research. Review articles, commentaries, perspectives, and reviews of existing protocols are excluded.
- The preprint, not the journal article, becomes the official record for HHMI’s own purposes. For Investigator reviews, Hrabowski Scholar midpoint evaluations, and Janelia Group Leader reviews, the most recent preprint version generally serves as the article of record.
- A budget consequence. For articles submitted on or after January 1, 2026, HHMI lab budgets can no longer be used to pay article processing charges (APCs) at hybrid journals; APCs at fully open-access journals can still be paid from HHMI funds.
Authors must also notify the journal (or ensure a collaborator does) that the article is subject to the policy, and inform collaborators of the requirement when a paper involves a major contribution from an HHMI lab.
What the Gates Foundation’s 2025 policy requires
The Gates Foundation’s policy applies to manuscripts arising from grants it funds and, in its refreshed form, took effect for funded manuscripts as of January 1, 2025. Rather than a flat “post before you submit” rule, the policy language is: funded manuscripts “shall be published as a preprint” on a foundation-recognized server “as soon as possible and to the extent feasible,” and “prior to formal peer review” — with a narrow exception where a grantee determines a preprint is inappropriate on ethical, safety, or other legitimate grounds. See the full Gates Foundation Open Access Policy guide on CASRAI for the full detail, including its CC BY licensing requirement and its separate, much-discussed decision to stop funding open-access publication fees (APCs) at subscription and hybrid journals, which pushed the no-cost preprint route to the center of compliance for Gates grantees.
Where the “before submission” shorthand breaks down
Neither funder’s actual text requires a preprint to exist at the literal moment a manuscript is uploaded to a journal’s submission portal:
- HHMI’s hard deadline is “sent for peer review,” not “submitted.” Submission-time preprinting is a recommendation, not the enforced trigger.
- Gates uses “prior to formal peer review” and “as soon as possible and to the extent feasible,” language that gives grantees room to sequence preprint deposit around a specific journal’s workflow rather than mandating it precede the submission click.
For research-administration purposes, the practical effect is nearly identical either way: by the time a manuscript is under peer review at a journal, a compliant preprint needs to exist. Advising authors to preprint at or before submission remains the safest operational guidance, because it satisfies both the recommended and the required reading of each policy and avoids any risk of a late preprint slipping past the point a journal sends the manuscript out for review.
HHMI vs. Gates Foundation: side-by-side
| HHMI (Immediate Access to Research) | Gates Foundation (2025 Open Access Policy) | |
|---|---|---|
| Effective for | Articles submitted on/after Jan 1, 2026 | Funded manuscripts as of Jan 1, 2025 |
| Who it covers | HHMI-employed scientists (Investigators, Janelia staff, named fellow/scholar categories) with a major-contribution authorship role | Grantees producing manuscripts from Gates-funded research |
| Hard timing requirement | Promptly once the journal sends the manuscript to peer review (submission-time deposit is recommended, not required) | Prior to formal peer review, “as soon as possible and to the extent feasible” |
| Second preprint required? | Yes — a revised preprint incorporating peer-review feedback, before journal publication | Not specified as a separate mandatory step |
| Required license | CC BY 4.0 only | CC BY (see full Gates guide for detail) |
| Designated servers | bioRxiv preferred; medRxiv, ChemRxiv, arXiv permitted by field | Foundation-recognized/preapproved preprint servers |
| Exceptions | None specified for the preprint requirement itself | Ethical, safety, or other legitimate concerns, at grantee discretion |
| Funding-side change | No hybrid-journal APCs from HHMI budgets for articles submitted on/after Jan 1, 2026 | Foundation ended funding of journal APCs generally, tied to the same 2025 refresh |
What changed for authors in practice
- Preprinting moves from optional-and-encouraged to a funder compliance obligation for anyone in a covered role at HHMI, or funded by a Gates Foundation grant subject to the policy — not just a personal choice about early dissemination.
- License choice is no longer negotiable. Authors used to defaulting to a preprint server’s most restrictive available license (e.g. CC BY-NC) now need CC BY on the preprint itself for HHMI-covered work, which has downstream implications for text- and data-mining reuse and for how the preprint can be cited and redistributed.
- The workflow gains a second checkpoint. HHMI authors now need to track a revised-preprint deposit as a distinct step between “peer review feedback received” and “journal publication,” not just a one-time preprint-and-forget action.
- Budget planning shifts. Labs that routinely paid hybrid-journal APCs from grant or institutional funds to secure immediate open access now need a different plan for those venues, since neither funder’s budget will cover that route the same way going forward.
- Journal notification becomes a required step for HHMI-covered authors, not just good practice — the policy explicitly requires notifying the journal (or ensuring a co-author does) that the submission is subject to the policy.
Frequently asked questions
Does the HHMI policy apply to every co-author on a paper, or just the HHMI scientist?
The policy attaches to the paper when an HHMI scientist subject to the policy, or a member of their lab, holds a “major contribution” authorship position — co-first, co-last, or co-corresponding author. HHMI scientists are responsible for ensuring compliance and for informing collaborators when their lab’s contribution triggers the requirement, even if the corresponding author is at a different institution.
Do review articles or commentaries need to be preprinted?
No. HHMI’s policy is scoped to “original research articles” describing original findings or novel protocols and methods; review articles, perspectives, and commentaries are explicitly out of scope.
What happens if a compliant preprint isn’t deposited in time?
Neither funder’s public policy text specifies a standard financial penalty for a single missed deposit in the way, for example, NIH’s Public Access Policy compliance can affect future funding actions. Research offices should still treat both policies as binding grant/employment terms rather than guidance, since HHMI ties the preprint to internal evaluation processes and Gates ties it to its funded-manuscript compliance framework; check current guidance from each funder directly, since enforcement mechanics can be updated independently of the underlying policy text.
Can a Gates-funded author still publish in a subscription journal?
Yes — the Gates policy governs how the manuscript is made openly available (via a compliant preprint, generally under a CC BY license) rather than which journal an author may ultimately publish in. See the full Gates Foundation Open Access Policy guide for how that interacts with journal choice and rights retention.
Related CASRAI coverage
- Gates Foundation Open Access Policy (2025) — full guide to the foundation’s policy
- Preprint Servers Explained — what preprint servers are and how they work
- Rights Retention Strategy — the funder-vs-publisher tension behind these mandates
- Publish, Review, Curate (PRC) — ASAPbio’s framework for preprint-first publishing
- Preprint Version Numbers — how initial and revised preprints get cited and tracked
