The adoption gap
Computing power is only the beginning.
Greece’s next AI advantage depends on the ability to turn infrastructure into repeatable operating capability.
Read the denominator
Eurostat’s 2025 survey records AI use in 8.9% of Greek enterprises, against 20.0% across the EU. It covers enterprises with at least ten people employed in the surveyed activities. It is not a census of every Greek business, and it does not measure the return on an AI investment. Treating an adoption rate as a productivity result skips the most important question: what changed in the work?
The missing middle
DAEDALUS adds a major national computing asset, as documented by EuroHPC JU. Our analysis is that the scarce organisational capability sits between access to technology and a dependable service: preparing data, redesigning a workflow, testing exceptions, training users and assigning responsibility. A faster model cannot repair an unclear process or create permissions to use data.
Why size changes the route
The Greek survey figures differ sharply by enterprise size: 7.3% for small firms with 10–49 people employed and 37.6% for firms with 250 or more. This association does not establish why the gap exists. It does make one uniform adoption programme a weak starting assumption. A smaller organisation may need a supported service, shared expertise and a bounded cost before it needs a larger technology stack.
An operating agenda
Choose one process with an accountable owner. Measure its current quality and effort. Test the full workflow, including review and failure handling, on a representative sample. Then ask whether the observed improvement survives a second team, a different period and ordinary operating pressure. Record the costs that remain after the demonstration: integration, supervision, licences, training and exit. The national question becomes concrete when more organisations can repeat this cycle without starting from zero.
Turn the argument into a decision.
Compare your next projectsSources & editorial method
Public sources establish context. The interpretation, questions and proposed exercises are AIStrategy editorial analysis. No participant, completed trial or outcome is implied. Reviewed 7 November 2026.