The Gap Is Not In The Technology

Start with the size of the room, because it is printed. The EIC Accelerator holds EUR 634 million for 2026. Of that, EUR 414 million sits in the Open call and EUR 220 million in the thematic Challenges. A grant runs below EUR 2.5 million at a 70 per cent funding rate. Equity from the EIC Fund starts at EUR 0.5 million and reaches EUR 10 million. Blended finance is the default, and around 85 per cent of recent winners took grant and equity together. Eligibility is narrow and clean. One small company or start-up applies on its own, at technology readiness level 5 or 6, moving toward level 8. There is no consortium. No partner search. No negotiation across fifteen organisations about who holds which work package. Six cut-offs sit in 2026: 7 January, 4 March, 6 May, 8 July, 2 September, 4 November. Two of the first three were evaluated together. That is the whole frame. Now the constraint nobody prices correctly. The deadline is not the obstacle. Twelve pages is. The programme does not want your ambition. It wants your arithmetic.

Step 1 is the short proposal, and every piece of it is capped. Part B runs to 12 pages maximum. The pitch deck is a PDF of 10 slides maximum, with no fixed template. The video is 3 minutes maximum in MP4, up to three team members, the chief executive recommended on camera. The national contact point data-sharing consent is mandatory and is routinely forgotten. Over length is not returned for trimming. Excess pages are removed before anyone reads them. Formatting is prescriptive to the millimetre: Times New Roman or equivalent at 11 point minimum, A4 pages, 15 millimetre margins minimum, single line spacing at minimum. Here is the part worth copying into your own working file. Each of the three sections maps to exactly one evaluation criterion. Section 1, technology, carries the deep-tech claim, the novelty, the evidence that TRL 5 is complete and the intellectual property strategy. It is scored on Excellence. Section 2, market, carries the total addressable market, the route to the customer and the value delivered. It is scored on Impact. Section 3, team and financial, carries the team, the skills gap each hire closes, investor traction and the funding rationale. It is scored on Implementation. So twelve pages is not a summary. It is three answers to three exam questions, with the weights printed on the paper. Four independent experts then return a GO or a NO GO.

The full proposal is scored from zero to 5 on each of three criteria. Each criterion carries 33.3 per cent of the weight. The thresholds are published: a minimum of 4 out of 5 on each one, and 13 out of 15 in total. Run the arithmetic, because it decides the shape of the document. A 5, 5, 5 scores 15. A 5, 5, 4 scores 14 and passes. A 5, 4, 4 scores 13 and passes on the last point available. A 5, 5, 3 also totals 13, and still fails, because no single box is allowed below 4. Two points of slack. That is the entire margin. A reactor with 1,300 validation hours fills one of those three boxes. The other two are written in prose about money and about people, and they are read with the same pen, on the same sheet, in the same hour. The container shrank as well. The full proposal is capped at 20 pages, down from 50. Nobody cut 30 pages of material by accident, so choosing what survives is a decision rather than a formatting task.

The programme reports its own arithmetic, round by round. The October 2024 cut-off drew 1,211 proposals. Of those, 431 were invited to jury interviews and 71 companies were selected. Of the 71, 79 per cent took blended finance, ten took grant only and five took equity only. A further 330 applications were assessed positively, were not funded, and went home with a Seal of Excellence. The October 2025 cut-off drew 923 full applications. Of those, 121 reached interview and 61 were funded. That is about 6.6 per cent of full applications, and nearer 3 per cent end to end once Step 1 is counted in. There is a strike rule on top of the queue. Three unsuccessful submissions at any stage and the door closes for the rest of Horizon Europe. The counter was reset to zero on 1 January 2024, so every applicant is working from a clean, finite count. The probability that a technically excellent dossier loses on its market and team chapters is not small. Most founders price it at zero. Most rejections are not bad ideas. They are good ideas described vaguely, with a market size asserted rather than evidenced and an intellectual property position that does not survive the first question.

A perfect reactor scores one third. The other two thirds are a document.

Step 2 is where a technical dossier becomes an investment case. The business plan, the implementation plan and the financial projections all have to fit inside the 20-page cap. Step 3 is the interview, and its shape is public. A 45-minute session built as a 10-minute pitch and 35 minutes of questions, in front of the EIC jury with programme managers and EIC Fund representatives in the room. An invitation goes to proposals worth roughly two to two and a half times the money available. Everyone else holds a Seal of Excellence and no cash. Timing is published as well. Step 1 returns in roughly four to six weeks. Step 2 takes two to three months. The January and March 2026 cut-offs were evaluated together, with results announced on 15 June 2026. The money arrives in tranches and not in advance, which is a cash-flow fact that belongs in the financial plan rather than in a surprise twelve months later. And the cheapest lever in the whole process is free. Every country runs a government-funded national contact point that gives independent advice on eligibility, the platform and the evaluation, at no charge, to anyone who asks.

By the time a score exists, the order of reading is fixed. The market size comes first, then the business model, then the team and the milestones, and only after that does the physics get its turn on the sheet. The failure modes are countable, and they repeat. A market size asserted with no named source. An intellectual property position with no filing month and no named office. A financial plan that contains no multiplication. A team chapter that lists qualifications instead of naming the gap each person closes. Every one of those is a blank in a scored box. Sixteen years of field work, from rural Nigeria to post-conflict Serbia to Kazan to Fukushima to Silicon Valley, and a EUR 75 million industrial group restructured along the way, taught the same lesson the programme prints in its own thresholds. Entrade built 200 wood-gasification machines across the UK and Europe. The machines were never the problem. The paperwork was. Write the three boxes and the technology stops being the argument you are defending. It becomes one third of the document you are submitting.