Estonia plans to buy €20 million worth of AI computing capacity

The total volume and strategic goal of the newly announced procurement by the European High Performance Computing Joint Undertaking (EuroHPC) are large: up to seven artificial intelligence gigafactories are planned to be built across Europe, from which the member states participating in the joint procurement and the European Commission will purchase services worth up to €10 billion.

In addition to Estonia, 17 other European Union member states that have signed a joint procurement agreement are participating in the procurement, the Estonian Ministry of Justice and Digital Affairs announced. Specially created consortia are invited to submit bids for the procurement, which can unite companies, public sector institutions, investors and other partners.

Thus, Estonian companies can also form consortia and submit their offers directly to the European High Performance Computing Joint Undertaking via the eSubmission digital system.

The successful consortiums will be jointly procured and guaranteed access to the necessary computing capacity by the participating countries and the European Commission.

The factories to be created do not necessarily have to be located in one country, but can also consist of cross-border and distributed artificial intelligence infrastructure. The Estonian government, through the Ministry of Justice and Digital, is ready to purchase up to 20 million euros in total of artificial intelligence high performance computing services from this consortium, which, if successful in the procurement, will establish a branch of a new generation artificial intelligence data center in Estonia.

The deadline for submitting bids is 12 November 2026. The contracts are planned to be signed in early 2027, and the first AI data centers should start operating no later than 18 months after the contracts are signed.

The European Commission's procurement aims to strengthen Europe's technological sovereignty and ensure that local startups, large companies, researchers and the public sector have access to the infrastructure needed, among other things, for training and processing large AI language models.