EU's Agile pilot adopted by large majority: funding decisions within four months for AI and drones
The European Parliament has given the green light to the EU's new defence innovation programme Agile: a one-year pilot programme of €115 million intended to help European startups and mid-sized companies bring promising defence technology to operational readiness more quickly. The decision was made with 536 votes in favour, 101 against and 16 abstentions on Tuesday, according to Heise online (source ID: 8c7714b7-12db-4199-9d70-fc2cd3acde3b). At the same time, parliament drew a clear ethical line: funding for lethal autonomous weapons systems without effective human control is excluded from the programme. But before any of this becomes money, formal approval from the Council remains – meaning the planned launch on 1 January 2027 is not yet assured.
What Parliament Approved
In Tuesday's vote, parliament adopted the regulation establishing Agile, a pilot programme intended to run for one year. The programme specifically targets startups and mid-sized companies – that is, the actors that often have the technology but lack the capital and contracting pathways to reach operational readiness in the defence sector. The purpose is to shorten the path from promising technology to actual use, Heise online writes, citing the European Parliament.
The vote result – 536 in favour, 101 against, 16 abstentions – shows a clear majority across group lines, although an opposition of 101 votes indicates the programme was not entirely uncontroversial. The sources do not provide information on how the groups voted, and no such claim should be inferred here.
The Ethical Line
The most notable aspect of the decision is the explicit exclusion: no funds from Agile may go to lethal autonomous weapons systems that lack effective human control. At a time when European defence investment is rising sharply, it is unusual for a funding programme to impose such a concrete stop on an entire category of technology.
The Greens' negotiator in the industry committee, Sergey Lagodinsky, framed the exclusion as a matter of rules: "We need clear rules for the use of new technologies and effective human control," he said, according to Heise online.
However, the exclusion raises an open question the sources do not answer: what actually constitutes "effective human control" is not defined in the available coverage. It is therefore unclear how strict the test will be in practice, and how it will be enforced in application and oversight processes.
How the Programme Works
Agile differs from much other EU funding on several points designed to make the programme faster and more accessible to smaller actors:
A four-month decision deadline. For key technologies such as quantum computing, AI-supported reconnaissance, cybersecurity and robotics, decisions are to be made within four months. For comparison, typical EU framework programmes often have considerably longer processing times – but the exact comparison figures are not part of the source base, and Agile is, after all, a pilot lasting only one year.
Individual applications. Companies can apply alone, without forming consortia. This removes a well-known threshold for smaller defence startups, which have traditionally had to band together with larger industrial partners to gain access to European funds.
Budget and duration. The pilot programme has €115 million available over one year. The sources do not specify how the amount is to be distributed across the technology areas, and hypothetical allocations would be pure speculation.
The Ukraine Component
A central element of the pilot is cooperation with Ukraine. Startups that receive funding are expected to collaborate with innovative Ukrainian companies that have accumulated extensive practical experience with digital systems in the ongoing war. This means war experience – particularly with drones, electronic warfare and digital command systems – is to be fed into European supply chains.
The rapporteur on the committee for security and defence, Tonino Picula, tied the two threads together: "The programme should not only bring defence innovations faster from concept to deployment, but also help Ukraine reduce strategic dependencies and increase supply security," he said, according to Heise online.
The wording points to a two-sided logic: Europe gains access to field-tested technology and knowledge transfer, while Ukraine is strengthened as an industrial partner rather than standing alone. However, the sources do not specify how the cooperation is to be organised in practice, or what legal framework applies to Ukrainian participants.
What Remains Before the Money Flows
Parliament's decision is an important step, but not the last. Formal approval from the Council of Ministers was still outstanding at the time of publication. Only once the Council has approved the regulation can Agile launch as planned on 1 January 2027 – and until then, the launch date is not guaranteed.
There are several open questions the sources do not answer but which will be decisive for the programme in practice:
- How will the €115 million be allocated across quantum technology, AI reconnaissance, cybersecurity and robotics? No published distribution exists in the sources.
- How will "effective human control" be defined concretely, and who will assess it in individual cases?
- Will the exclusion of autonomous killer weapons carry over into the larger EU funding instruments for which Agile is described as a blueprint? That is unclear, and the outcome will determine whether the ethical line becomes a one-off or a durable standard.
Why This Matters
In scale, Agile is a modest programme – €115 million is small in the context of European defence budgets taken as a whole. The significance lies in the design. The programme tests a new model for funding defence innovation: fast processing, individual applications and direct access for small actors, combined with an explicit ethical constraint. If the pilot works, it becomes a template for later, larger instruments – and both the speed model and the exclusion of autonomous killer weapons will then matter far beyond this one-year allocation.
For now, two things remain to be seen: the Council's approval, and how the definition of "effective human control" translates into actual funding decisions once the first applications arrive.