The popular story goes like this: the European Research Council pays for excellent science, and the scientist then decides where to publish. The story is wrong. An ERC grant now comes with a publication mandate precise enough that the real decision is not whether to make a paper open; it is whether your author accepted manuscript reaches a trusted repository before your project officer asks for it. That operational detail, not the rhetoric about open science, is what researchers feel first.

Under the Horizon Europe programme that started in 2021, ERC grants follow the European Commission's open science requirements, which are stricter than the Horizon 2020 rules they replaced. For new grants, the default is not an embargo of six or twelve months. The default is immediate open access under a Creative Commons Attribution licence, commonly shortened to CC BY. The shift catches researchers who won their grant under one set of rules and are now reporting under another.

The compliance clock starts earlier than most researchers think

Compliance under Horizon Europe is date-driven in a way that trips up even experienced grantees. The grant agreement does not say "eventually." It says deposit at the latest at the time of publication. That means the repository entry and machine-readable copy of the published version or final peer-reviewed manuscript must exist when the journal article goes live, not three months later when someone in the faculty office processes the files.

For Horizon 2020 grants, the clock was softer: full-text open access within six months of publication, or twelve months in the social sciences and humanities. Researchers moving between grant generations often quote the old embargo windows as if they still applied. They do not.

Confusion is not just an administrative nuisance. Project officers can record open access failures as compliance issues, and unresolved obligations have consequences under the grant agreement. The real risk is rarely a dramatic legal penalty; it is the slow, bureaucratic consequence of a flagged deliverable that nobody on the team remembers signing up for.

What the grant agreement actually says

The Horizon Europe model grant agreement contains the specific clauses. Beneficiaries must ensure open access to all peer-reviewed scientific publications relating to their results. The deposit must happen in a trusted repository at the latest at the time of publication, and the deposited file must carry a CC BY licence or one with equivalent rights.

The authors must retain sufficient intellectual property rights to comply. This rights retention point is the part publishers least like and researchers most often miss.

The table below compares the two regimes because grantees regularly ask where the old six-month option went.

One point deserves underlining: the published version is not required. The final peer-reviewed manuscript accepted for publication, sometimes called the author accepted manuscript, is sufficient. The ERC does not require you to pay for gold open access.

Here's the catch

Immediate open access is not a publishing choice; it is a retained right. Many publishers will still hand you a copyright transfer form that asks for exclusive rights that conflict with your grant agreement. Sign it unamended and you can be in breach before the article gets a DOI. The grant agreement does not suspend itself because a publisher's workflow is from 2009.

Rights retention operates through a pre-publication addendum, often provided by university libraries. It states that the author retains the right to deposit the accepted manuscript under CC BY immediately, regardless of publisher policy. The cOAlition S Rights Retention Strategy is the reference point many institutions have adopted.

Some publishers respond by refusing to handle a paper with such an addendum; others accept it silently. The researcher's job is not to win an argument with the publisher. It is to make sure the signed agreement does not strip away what the grant requires.

This is the point where the mandate stops being a distant policy ideal and becomes a form on your desk.

Repositories, rights retention, and the publisher form

A trusted repository is not any website. The Commission expects an institutional repository, a subject repository such as arXiv or Europe PMC, or a certified service. Many institutions run OpenAIRE-compatible repositories; CoreTrustSeal certification is the standard you can point to when someone asks what "trusted" means. OpenAIRE maintains guidance on compatible services.

Deposit the author accepted manuscript, not the publisher's typeset PDF. Include the grant ID, project acronym, title, authors, licence, and a link to the published version. The metadata should be open under CC0 so discovery systems can index it properly.

For publisher policy checks, Sherpa Romeo lists journals and their self-archiving policies. It is not a determinant of your rights, because your grant and the rights retention statement override the publisher's boilerplate, but it helps you anticipate friction. As AcademicJobs has noted, journal indexation is not a quality guarantee.

What this costs, and who gets invoiced

Open access mandates have a habit of turning into invoices. The ERC grant can cover publication fees, but only for fully open access venues. Hybrid journals, the ones that charge readers and authors at the same time, are not eligible for APC reimbursement under Horizon Europe. That one sentence kills a lot of budgets.

There is no separate ERC open access fund. Publication fees come out of the project's direct costs, competing with reagents, travel, and salaries. A EUR 3,000 article processing charge is a centrifuge rotor you do not buy. Researchers should build publication costs into the proposal or confirm their institution has a library agreement that covers them.

Some universities have read-and-publish agreements with major publishers that cover APCs for corresponding authors. Those deals sometimes require the journal to be hybrid or fully open access within a specific package. The institution's library can confirm what is already prepaid before you choose a journal. The European Commission's open access policy pages set out the finance rules for full open access venues.

Hype versus reality

The mandate's cheerleaders say it makes tax-funded research available to anyone with a browser. That part is true, but not because publishers chose it. The repository deposit does the heavy lifting while paywalled versions still generate subscription revenue.

The mandate's critics say it punishes researchers who cannot pay APCs. The grant agreement's green route answers that: you do not need to pay an APC to comply. The accepted manuscript can go into the repository, immediately, under CC BY. No credit-card step.

What it does not do is dismantle prestige publishing. A CC BY deposit does not change where a Nature paper is submitted or how hiring panels weigh it. The mandate changes accessibility, not the reward structure. Anyone who tells you otherwise is selling something.

Hype: everyone will now read taxpayer-funded science. Reality: the file is available; discoverability, citation culture, and search habits still determine who actually opens it.

A compliance checklist that works

The mandate is only as burdensome as the workflow around it. The researchers who find it manageable tend to do five things in the same order every time.

• Confirm whether the grant was signed under Horizon 2020 or Horizon Europe, because the rules differ.
• Keep the author accepted manuscript from the moment of acceptance, before any publisher PDF replaces it.
• Check the publisher agreement for clauses that conflict with rights retention, and attach an author addendum if needed.
• Deposit in a trusted repository at the time of publication with the grant ID, CC BY licence, and full metadata.
• Record the repository link in the project reporting portal so the compliance trail is visible to the project officer.

For a postdoc on an ERC team, the mandate means one less thing to negotiate: the accepted manuscript goes into the repository whether the publisher likes it or not. That is the point.

Photo by Tim Mossholder on Unsplash