The Innovation Act over-promised and over-delivered

When the Commission unveiled its Startup and Scaleup Strategy in May 2025, I argued it proposed some useful supply-side measures (more venture capital, easier access to talent, less fragmented regulation) but startups still faced a demand-side dilemma: namely, the Strategy didn’t help create customers to buy their innovation.

This was backed up by data from founders. Slush’s 2025 Startup Struggle Survey found that around half ranked customer acquisition as a top-three concern. The share citing revenue growth as a top concern jumped from 36.8% in 2024 to 52.8% in 2025.

Fortunately, Europe’s innovation community remains vocal. The March 2026 European Deep Tech Reportcalled on policymakers to “shift governments to become customers” by involving startups in procurement, reforming vendor vetting and prioritising commercial contracts over grants.

Enter today’s European Commission proposal for an Innovation Act.

The Commission has – I think it’s fair to say – had its own Go-To-Market issues with the Act. An internal policy audit had to send it back for revisions twice before it finally got put on the docket. A lot of the original ideas around including regulatory sandboxes, mandatory targets for innovative procurement, and other far-reaching horizontal measures got cut, and the final proposal was drastically slimmed down.

I believe this nets out positive. I have argued the Commission needs to get better at identifying the single, white-hot constraint innovators face in deploying innovation, and focus their efforts on removing it. The Slush data shows that, for founders, one of the top constraints is finding customers to generate early revenue.

And this, to its credit, is exactly what the Innovation Act aims to do. The draft creates a new framework for public authorities to procure R&D and pre-commercial development, effectively paying innovators to develop, test and demonstrate the unproven technologies a government needs, and rewarding them for reaching development milestones. (We aren't talking about finished products here: the Act stops before commercial-scale purchasing.)

The measures also seem to explicitly target the sort of problems a startup and scaleup would face along the procurement cycle.

1.  By the time startups and scaleups see the tender it is too late: Public buyers would have to consult the market before launching the process and publish that consultation EU-wide, giving scaleups a chance to demonstrate what they can deliver before the tender has effectively decided what needs to be procured.

2. The tender was designed for the incumbent: Buyers would generally have to specify the functional requirements they need, rather than a particular design, process or technological solution, giving challengers more room to compete with genuinely different technologies.

3. Startups don't have the track record to qualify: Buyers could not normally require previous contract experience, while startups and scaleups could demonstrate financial capacity through business plans, IP, investor funding or their own capital. Minimum financial-capacity requirement would be capped at 50% of the contract value, ans guarantees would be halved for qualifying startups and scaleups.

4. Government procurement incentivises risk-avoidance over innovation: Europe directs 10% of public procurement towards innovation, versus 20% in the US and 25% in South Korea. A Commission Expert Group estimates closing the procurement gap could generate up to €300bn in new contracts. To realign incentives, under the Act, innovation would count for at least 15% of the tender score and quality at least 50%, with buyers also able to reward European technology, supply chains and economic activity.

5. Startups have to beat the incumbent outright: Buyers would generally have to select multiple suppliers, allowing competing technologies to be developed and tested through successive phases before the field is narrowed.

6. One national contract isn't enough to build a market (1): Public buyers from different Member States could jointly procure R&D, pooling their requirements and funding. Participation would generally be European or from countries offering reciprocal access, while at least half of the underlying R&D would have to take place within that geography. (2)

7. If government pays to develop it, what happens to the IP? Companies would generally retain ownership of IP they create, so they can commercialise the resulting technology elsewhere. While buyers would retain access rights and safeguards for  secure licenses or transfers where the technology is not commercialised, or to protect “strategic interests.”

So, does it solve Europe’s demand-side dilemma?

Not quite. But it does a decent job attacking part of the problem: removing several of the barriers that stop an innovative company reaching a first public customer.

The Act alone won't turn risk-averse procurement teams into venture customers, and much will depend on how aggressively public buyers actually use the new framework, but the measures above should make it easier for startups and scaleups to get in front of public buyers, qualify for tenders, compete on the strength of their technology, and get paid to develop it.

The problem, then, is what happens next. The Act is deliberately limited to R&D and pre-commercial development. It can help a company develop, test and demonstrate a technology, but it does not solve the problem of getting governments to buy that technology repeatedly, in commercial volumes. The Commission's own research suggests this is a major weakness: successful pilots still rarely translate into scaled solutions.

Public procurement is potentially an enormous source of demand. European public authorities spend around €2 trillion a year purchasing goods, services and works. The Innovation Act only touches a relatively narrow part of that.

For now, though, I think the proposal deserves credit. It was promised as a sprawling piece of innovation legislation and emerged as something much more limited. But if good policymaking means identifying the critical constraint and attacking it directly, narrower isn't necessarily worse. In overpromising, the Commission may have overdelivered.

The next question is what happens when the technology works, and whether government can move from being a first customer to becoming a market at scale. That is the benchmark the other proposal released today – The Public Procurement Act – should be measured against.

***

(1) 80% of European defence procurement, for example, remains national.Although, interestingly, defence procurement is notably outside the scope of the Act.

(2) I understand the logic of European-preference provisions to help home-grown challengers win their first European customers. But it remains to be seen whether their design could create complications once those companies raise from third-country investors or expand into markets outside Europe.

- Brendan Moran is Director at DELTA-V Public Affairs