EIC Accelerator for deeptech: what the 2026 scoring means for technical founders

The EIC Accelerator is the largest single non-dilutive cheque most European deeptech founders can realistically apply for. Up to two and a half million euros in grant, another five in equity, and a validation stamp that makes later conversations with customers and investors easier. It is also one of the most painful applications a technical team will ever write.

What changed for the 2026 calls is not the headline numbers. It is the weighting. The evaluators now score the commercial case and the team more aggressively, and they are visibly tired of technologies solving problems that have not been articulated as customer pain. A beautiful physics result with a vague market is no longer enough.

Read the call like an evaluator, not like an applicant

There are six criteria: excellence, impact, implementation, risk, team, and financials. Founders naturally gravitate to excellence because that is where the technology lives. But the criteria are scored independently, and a low score on impact or team is almost impossible to recover from, however good the science is.

Impact means one specific customer group, one measurable problem, and a credible path from the project result to revenue. It is not TAM. It is not "the energy market". It is a named buyer, a budget line, and evidence that the buyer is already trying to solve this.

Team means the people who can run both the technical work and the commercial work. A founding team of two excellent scientists and no one who has spoken to a customer will score poorly. A team with one founder who can hold a customer conversation and another who can hold the technical roadmap will score well, even if the technology is earlier.

The proposal is a story, not a report

The most common mistake I see is treating the application as a dump of every result the lab has produced. Evaluators read dozens of these. They are looking for a coherent arc: here is a hard problem, here is why existing solutions fail, here is our unique answer, here is the customer who pays for it, and here is exactly what this grant buys.

Every figure, every TRL claim, every competitor mention should serve that arc. If it does not, it is noise. The page limits make the discipline easy to enforce: if a paragraph cannot be traced back to one of the scoring criteria, delete it.

Budget realism is a scoring item in disguise

The financials section is not just a spreadsheet. It is a test of whether the founders understand what it actually takes to get from here to there. Costs that are too low read as naive. Costs that are unexplained read as padded. The safest budget is one where each work package has a clear output, a clear owner, and a clear link to the impact story.

Do not forget the equity question. The Accelerator now forces a go or no-go on the equity option early in the process. Founders who have not modelled dilution before the interview are easy to spot and easy to reject.

The interview is where most good proposals die

If you reach the interview stage, the written score is no longer the deciding factor. The jury is testing whether the team can defend the project under pressure, whether the founder can speak the customer's language, and whether the technology claim survives skeptical questions.

The teams that pass are the ones who have rehearsed the hard questions: what happens if the key technical assumption is wrong, who is the first customer, why does this not end up as a feature of a larger platform, and what does the balance sheet look like the month after the grant ends.

Should you apply?

Apply if the project fits the call, the team can spare one senior person for most of three months, and the non-dilutive money changes what is commercially possible in the next eighteen months. Do not apply if the main motivation is validation or runway. There are cheaper ways to get both.

If you are thinking about the next round of EIC Accelerator and want the commercial story pressure-tested before the writing starts, this is the work we do on the grant side.