Digital Europe programme – interim evaluation

European Commission · Have your Say · Initiative 13880

Interim evaluation of the DIGITAL programme (2021–2027) – Report COM(2025)771 · Statement of October 28, 2026

Core statement: An evaluation that measures expenditure and disbursements does not measure what remains. The report counts computers, hubs, trainees and companies – but the decisive question is: which structure stands after the funding? Structure in the sense of Trusted WEB 4.0 is European-controlled, decentrally resilient and accessible to all through participation – and the Digital Europe programme must be measured against these three yardsticks, not against the sum of its funding decisions.

What the report finds

The interim evaluation of the Digital Europe programme (Report COM(2025)771 of 18 December 2025, Staff Working Document SWD(2025)420, 316 pages) documents four years of implementation: financing of JUPITER, Europe’s first exascale supercomputer (rank 4 in the TOP500); four large-scale AI Testing and Experimentation Facilities; more than 150 European Digital Innovation Hubs covering some 90% of European regions, which have delivered more than 29,000 services to over 54,600 companies; more than 20,700 people trained in highly specialised digital fields; 27 national coordination centres in cybersecurity; and investments in semiconductor pilot lines and competence centres. The call for evidence and public consultation ran from 27 June to 20 September 2024 and received 103 contributions – among them the clear citizens‘ demand to make open source the basis of all funded technology development.

Recognition: broad reach and European infrastructure

GISAD expressly acknowledges two design features of the programme. First, the decentralised footprint: innovation hubs in 90% of the regions and national coordination centres instead of a single central contact point – many nodes instead of one centre is the right administrative logic. Second, European ownership of strategic infrastructure: a European exascale computer is an asset, not a rental – and every infrastructure that Europe owns instead of leasing reduces dependence on US hyperscalers and the CLOUD Act.

Criticism 1: capacity funding without an origin yardstick

The report measures what was funded – not in whose hands the funded lies. The question GISAD already raised for the Digital Decade adoption targets applies all the more to the billions in funding: how much of the funded technology, the hub services and the AI infrastructure runs on European-controlled Stage-B infrastructure – and how much on US platforms subject to the CLOUD Act? Funding that co-finances US dependence reinforces the very problem it is meant to solve. Every funded project should report an origin yardstick: owner, operator, legal order of the core systems. What is not European-controlled cannot be counted as European digital capacity.

Criticism 2: exascale lighthouses are no substitute for resilience

JUPITER is a legitimate success – but bandwidth is not resilience. Decentralised digitalisation does not replace a single cable – but it takes the catastrophic effect away from every cable break. A programme that concentrates the lion’s share of its computing-power policy in a few giant nodes reinforces precisely the concentration whose failure the recent years have demonstrated – from CrowdStrike to the Charité to the Baltic Sea cables. Resilience is created by many medium-sized, networked, European-controlled nodes. The lighthouse belongs in the portfolio – but the yardstick must be the decentralised network, not the peak performance of one site.

Criticism 3: qualification as specialism instead of participation

More than 20,700 people trained in highly specialised fields is good – but with this the programme measures the top, not the breadth. Digital participation in Europe means: the elderly keep up. People in minority languages keep up. Those who do not want to search but want to be found (getmySense thinks in 2,500 written languages) belong in the measurement. Participation instead of welfare means, for a funding programme: digital participation infrastructure is not a line item of social systems, but a funding strand in its own right – measured by usage, not by course numbers.

GISAD’s recommendations

1. Origin yardstick as an evaluation indicator: every funded project reports ownership and operating arrangements as well as the legal order of its core systems; only European-controlled infrastructure (Stage-B criterion) counts as European digital capacity.
2. Resilience as a yardstick instead of peak performance: a funding quota for decentrally networked, medium-sized nodes; the exascale lighthouse remains a supplement, not the measure.
3. Open source as a funding condition – the citizens‘ contributions to the consultation demanded it, and the logic is right: what Europe funds, Europe must be able to develop further afterwards.
4. A participation strand with its own indicators: usage figures of the elderly, linguistic diversity of the offerings, accessibility without compulsory identification (WAN-anonymous where identity is not needed).
5. For the AI Testing and Experimentation Facilities, the yardstick of the AI safety debate applies: measurement outside the manipulable area, escape safeguard as a testing criterion, and sovereign operating environments – no US hyperscalers as the test framework for European AI.