Skip to content
Welcome, thoughtbot's Giant Robots listeners!Start here
European Tech Sovereignty
22OCT

Greece unveils DAEDALUS on Nvidia chips

2 min read
17:24UTC

Greece inaugurated the €36m DAEDALUS supercomputer at Lavrion on 19 October; nearly two-thirds of the bill came from EU recovery money spent through Athens.

← Back to EU questions AI labs over loss of controlJump to analysis ↓
TechnologyAssessed
Key takeaway

Greece's new AI supercomputer runs on Nvidia chips, paid mainly from EU recovery funds.

Greece inaugurated DAEDALUS on 19 October at Lavrion, in a data centre built inside a former power station, with Prime Minister Kyriakos Mitsotakis present⁠1. HPE built it on more than 2,000 Nvidia GH200 superchips, which pair a processor and an AI chip on one board. It ranked 31st on the June TOP500, the twice-yearly ranking of the world's fastest supercomputers, and anchors PHAROS, the Greek AI Factory.

The machine's acquisition cost was €36m. EuroHPC JU, the EU body that co-buys supercomputers with member states, paid 35%, about €12.6m; the other 65%, about €23.4m, came from Greece 2.0, the national recovery plan financed by NextGenerationEU⁠2. Because that larger share is EU recovery money spent through Athens, it shows up in a national plan and carries no EuroHPC label. Cyprus, Montenegro and North Macedonia sit in the consortium, the last two outside the EU.

DAEDALUS joins LUMI-AI , MeluXina-AI and Sweden's Arrhenius on the list of EuroHPC-linked machines contracted or opened since July; none of the four runs on a European accelerator.

Deep Analysis

In plain English

A supercomputer is a very large computer made of thousands of linked processors. Scientists use them for weather forecasts, drug research and training AI. Greece has opened one called DAEDALUS inside an old power station at Lavrion, on the coast south-east of Athens. It uses chips from Nvidia, the American company that dominates AI chips. EuroHPC is an EU body that buys supercomputers jointly with member states. It paid about a third. The rest came from EU recovery money that Greece received after the pandemic and chose to spend on this machine.

Deep Analysis
Root Causes

EuroHPC co-funds machines but leaves siting, procurement and most money to the host country. That design lets each state choose quickly, and the fastest choice is the market-leading accelerator.

Recovery funds had to be spent by fixed deadlines. Greece 2.0 money suited a ready-to-deliver vendor system rather than a longer bet on European silicon.

No European accelerator is yet sold at the scale these machines need, so every recent EuroHPC-linked system, from LUMI-AI to MeluXina-AI, has gone to US chips.

What could happen next?
  • Opportunity

    Converted power stations may become a template for faster supercomputer siting in other member states.

    Medium term · Suggested
  • Consequence

    Including Montenegro and North Macedonia extends EU-funded compute to candidate countries outside the bloc.

    Short term · Assessed
  • Risk

    Each new Nvidia-based machine deepens the software lock-in that a future European accelerator would have to overcome.

    Long term · Suggested
Sources:EuroHPC JU
Mentions:Greece →EuroHPC JU →HPE →Nvidia →Kyriakos Mitsotakis →NextGenerationEU →Cyprus →North Macedonia →MeluXina-AI →Tokyo →Athens →China →Beijing →Japan →Sweden →California →Bruegel →
First Reported In

Update #17 · EU questions AI labs over loss of control

EuroHPC JU· 22 Oct 2026
Read original →
Causes and effects
Caused by
Third EuroHPC machine, no European chip
DAEDALUS follows Arrhenius as another EuroHPC-linked machine with no European accelerator.
Occurred 8 Sept 2026
Read story →
€80m buys European hands, US silicon
Like MeluXina-AI, DAEDALUS runs on Nvidia accelerators.
Occurred 4 Aug 2026
Read story →
This Event
Greece unveils DAEDALUS on Nvidia chips
Most of the money behind Greece's AI Factory is EU recovery funding booked in a national plan, and the machine itself rests on US chips.
Different Perspectives
European Centre for International Political Economy
European Centre for International Political Economy
ECIPE found nine states gave about 97% of awarded cloud lots to home providers from 2018 to 2025, and every sovereignty-framed lot went to a domestic supplier. The think tank reads sovereignty clauses as a home-bias tool that narrows competition, with a 46% sole-bidder rate on a small sample.
Mistral AI
Mistral AI
Mistral opened a Munich industrial AI hub on 28 September with BMW and Siemens Energy work, and Germany's digital minister Karsten Wildberger praised it on the launch page. The French lab gains a German ministerial endorsement while its 21 October Nvidia partnership ties the work to US accelerators.
AWS European Sovereign Cloud
AWS European Sovereign Cloud
AWS will run its Brandenburg region for several hours on 24 October cut off from the AWS global backbone, with an EU-resident team. AWS says the test follows BSI's C3A criteria and will give regulators and customers verifiable evidence of separation.
Greenland government
Greenland government
Greenland signed a joint declaration with the EU and Denmark in September and received Virkkunen on 24 and 25 September. The Arctic Strategy now backs a cable linking Greenland directly to Europe, an alternative to Greenland Connect, which lands in Canada and Iceland.
Greek government
Greek government
Kyriakos Mitsotakis inaugurated DAEDALUS at Lavrion on 19 October, a €36m machine on Nvidia GH200 chips that anchors the PHAROS AI Factory. Athens paid 65% through Greece 2.0, making national recovery money the main funder of an EU AI Factory.
ASML
ASML
ASML reported record quarterly sales of €11.66bn and said planned supply for 2027 and 2028 will not fully meet demand. It plans about 30% more capacity in 2027, and some customers want to book for 2029 and 2030.

Written with AI from Lowdown's published reporting. Not individually reviewed by an editor. Editorial standards.