What Happened to EIC Accelerator Winners: A Follow-Up on 831 Companies

About a year ago, when we started to use AI a lot, we had the idea of creating follow-ups for EIC Accelerator winners. Very often you see that a company was announced as a winner, or you see the list of funded companies, and you wonder what is going to happen one, two or three years later, because you never hear from them again.

The only thing you see is the official reports, such as the innovation report or the tech report, and they only highlight the companies that have managed to scale or to win. The ones that are insolvent or bankrupt are never mentioned again.

From One Article per Company to a Data Room

So we created automated follow-ups: every year, research looks at what happened to each of these companies and whether there are any updates. The first version was okay. It was basically one article per company, each trying to assess how that company was doing.

Now the whole thing has been revamped. The principle is the same, every single company is researched, but the result is much more structured. It is a data room on eiceu.com, and for every winner it shows very clearly: is the company still active? Has it managed to scale? Has it gone insolvent, dormant or bankrupt? So you get a real follow-up.

The 831 companies won 841 awards, because 10 companies won twice. Three of them won in June 2021 and again in 2024 or 2025: Sparrow Quantum, VarmX and CrannMed. Germany has the most winners with 123, followed by France with 122 and the Netherlands with 89, so these three countries account for 40.2% of all winners. 15.2% of winners come from outside the EU, and Israel, with 48, has more than Italy, with 30. The country pages on eiceu.com break this down further.

Why the Research Stops at 2024

The data set is limited to companies since 2021. In the data room on eiceu.com you see all winners from 2021 until the last cut-off, but the in-depth research currently runs only to 2024. You have to consider that these companies need some time to do something. If you do a follow-up one year after the award, not much has happened, so better data needs time.

So far, 630 companies (75.8%) have been checked in depth: every winner from 2021 to 2023, half of the 2024 winners, and none of the 2025 and 2026 winners yet. How each company is assessed is explained in the eiceu.com methodology.

What Surprised Us: How Many Have Already Failed

We expected a certain number of companies to go insolvent. What surprised us is that for the early cohorts, roughly 2021 and 2022, about ten percent have already gone under. A lot of these companies have very little staff. They are not going to hire 50, 100 or 200 people, because EIC Accelerator companies usually have very small teams. If you give them a grant of up to EUR 2.5 million, or potentially also an equity investment, the money should last for a very long time, unless they have very poor financial management.

So it was surprising to see how many companies have already gone insolvent. Right now that is 44 companies insolvent, 7.0% of the checked winners, plus companies that are dormant, which effectively means the same thing: the website is offline and, while they may not have formally closed the company, they have completely ceased operations. Counting every kind of failure, 53 companies (8.4%) have failed. The full list is on the eiceu.com insolvencies page.

Failures rarely come quickly. The median company fails 2.9 years after its award, and only 7 of 51 failures came within two years. The number of failures per year is still rising: 3 in 2023, 12 in 2024, 19 in 2025 and already 17 in 2026 by the end of September.

Not a Good or a Bad Sign

Importantly, we do not think this is a good or a bad sign. You cannot say that the EIC invested and ten percent of the companies ended up insolvent, so the programme failed. It is neither a good thing nor a bad thing, because the EIC is investing in high-risk companies, and of course you expect a bunch of them to fail.

The way the EIC does its due diligence is also not the way a bank would do it. There is a process of looking at the proposal, and then the jury has to make a decision pretty quickly, based on an interview and the proposal. There is no proper due diligence in the sense of an analyst looking at all your numbers and at the verification evidence before making a funding decision. That does not happen. The due diligence happens after the award, which is how the EIC Accelerator likes to do it, and that is simply how the bureaucracy works.

The risk is not spread evenly. Biotech is the riskiest sector, with 19.5% of its winners failed, followed by energy storage at 18.8%. No quantum computing company has failed, and 65.0% of quantum companies are scaling. None of the 20 photonics companies has failed either.

By country, Denmark has the highest failure rate at 20.8%, followed by Germany at 15.6% and France at 14.9%. Spain, Italy, Ireland and Portugal have no failures so far.

Whether a company delivered its EIC project is the clearest warning sign. 28.4% of projects were fully delivered, 44.2% partly, and 29 projects (4.5%) were not delivered at all. Among the projects that can already be judged, 36.8% were fully delivered. Of the companies that did not deliver, 69.0% have failed; of those that fully delivered, only 2.2% have. No company that failed to deliver its project is scaling. The delivery breakdown on eiceu.com lists each project.

What the Data Room Shows for Every Cut-Off

For every single cut-off, you can now see which company is still active, which company managed to scale and how many exits there were, meaning companies that were acquired or that had an IPO. It is all there, so you get a chart for every cut-off showing which company did what. The cohort view on eiceu.com compares the cut-offs side by side.

24.1% of the checked companies are clearly scaling, 22.1% clearly are not, and for 53.8% it is unclear. Companies backed by the EIC Fund do much better: 53.2% of them are scaling, over double the average. The scaling page on eiceu.com names them.

30 companies (4.8%) have been acquired, and the median acquisition came 2.8 years after the award. AI companies exit most often, with 14.1% of them already exited, and Finland has the highest exit rate by country at 16.1%. All exits are listed on the eiceu.com exits page.

Kept Up to Date With Every Cohort

The data room will be kept up to date: every time a new cohort comes out, and the research will be repeated regularly to make sure everything is accurate. Of course, parts of it will be stale for a few months or even a year before they are updated. Even so, this data room is going to be very interesting to follow over the next few years. You can explore it on eiceu.com.

Summary

• A data room on eiceu.com follows up on all 831 EIC Accelerator winners since June 2021, across 14 cut-offs and 31 countries.
• 630 companies (75.8%) have been checked in depth, covering every winner from 2021 to 2023 and half of the 2024 winners.
• 44 companies are insolvent and 53 (8.4%) have failed in total; 12.8% of 2021 winners and 10.1% of 2022 winners have failed, against 4.7% of 2023 winners and 1.5% of the 2024 winners checked so far.
• The median failure comes 2.9 years after the award, and failures per year rose from 3 in 2023 to 19 in 2025.
• This is not a bad sign in itself: the EIC funds high-risk companies, decides after a proposal and a short interview, and does its due diligence after the award.
• Biotech (19.5%) and energy storage (18.8%) fail most; quantum computing and photonics have no failures.
• 69.0% of companies that did not deliver their EIC project have failed, against 2.2% of those that fully delivered.
• 84.8% are still active, 24.1% are clearly scaling, 30 were acquired and 2 are listed.