Greece's Leader Says Governments Are Behind on AI

Greek Prime Minister Kyriakos Mitsotakis told a San Francisco audience of about 250 founders, investors and operators on September 22, 2026 that no government is ready for what AI will do next. He spoke with a TechCrunch editor at an event that combined open discussion of AI governance with a pitch for Greece as a home for tech companies and technical talent. TechCrunch

TechCrunch headlined the conversation “We're already fighting yesterday's battle.” The line captured his central admission. He said he does not have answers to many of the AI questions political leaders are now debating.

He was blunt. That directness set the tone for the rest.

Mitsotakis described the Bay Area visit as partly a fact-finding mission. He said he had toured Tesla and Sequoia Capital that morning before taking the stage, and planned to travel to New York later that week for the U.N. General Assembly.

The pitch focused on infrastructure and policy changes from recent years. Mitsotakis said Greece received about €36 billion from the EU's post-COVID recovery fund, with much of it directed to digital infrastructure. That means public services rebuilt around digital identity, or online proof of who you are, digital payments and government software running on cloud servers instead of paper files.

On computing power, Mitsotakis said Greece has assembled a new supercomputer with Hewlett Packard Enterprise in Lavrio that should come online within months for AI and scientific research. Computing power here means the processing capacity used to train and run AI models. He gave no system specifications or allocation policy in the remarks covered, and the timeline is still prospective.

On talent and company formation, he cited three domestic changes. Greece changed how stock options are taxed, loosened labor laws, and offers returning Greeks much lower taxes for up to seven years. Stock options, the right to buy company shares at a set price later, shape how startups hire, while labor rules shape early team structure and time-limited tax breaks shape senior hiring.

The broader context here is a shift in how smaller states compete for technical work. National computing capacity, digitized public administration and startup-friendly treatment of company shares were niche policy topics five years ago. They are now routine in investment-location discussions, alongside power supply, internet connectivity and immigration rules.

In my view, the tension is between slow-moving government and fast-moving AI. Mitsotakis framed policy as still arguing over earlier questions about data, platforms and jobs while the cost of running AI models falls and AI agents start doing real work inside companies. Lawmaking, procurement and testing move slowly. Model development does not, and that mismatch is structural rather than the fault of any single administration.

What is worth watching for technical leaders is the second half of his argument. A country that cannot train, test or run large models at home must rent that capacity elsewhere, with tradeoffs in speed, data control, cost and research independence. A national system in Lavrio will not remove those tradeoffs on its own. It does give Greek universities, labs and startups a local machine to experiment on and a stronger position than renting cloud access alone. Over time that kind of shared capacity compounds, as students trained on accessible machines become the engineers who build the next ones.