Lumi AI sera construit par Bull

EuroHPC has awarded Bull a €388 million contract to build LUMI-AI, due to come online in Finland in 2027. The machine should deliver a tenfold increase in AI capacity and gives the newly state-owned French company a major early test. But its architecture captures Europe’s central dilemma: more infrastructure on European soil, without independence from critical US technology suppliers.

Europe’s EuroHPC programme has awarded Bull a €388 million contract to build LUMI-AI, a new AI-optimised supercomputer to be installed in Kajaani, Finland. Bull said that it is the largest contract in the French company’s history.

The machine is scheduled for deployment in the second half of 2027 at CSC – IT Center for Science’s data centre, alongside the existing LUMI supercomputer. EuroHPC will cover half of the project through the Digital Europe Programme, with the remaining 50% financed by the LUMI AI Factory consortium of Finland, Czechia, Denmark, Estonia, Norway and Poland. The €388 million budget covers acquisition, delivery, installation and maintenance.

EuroHPC expects LUMI-AI to provide roughly ten times the AI capacity of the current LUMI system. Bull says conventional high-performance computing capacity will almost double. The much larger gain for AI reflects the increasing specialisation of modern processors for the lower-precision operations widely used to train and run neural networks.

European integration, American silicon

LUMI-AI will be based on Bull’s BullSequana XH3500 architecture, using AMD’s next-generation Instinct MI430X accelerators and sixth-generation EPYC processors with 256 cores. IBM will supply storage technology based on Storage Scale, while Finland’s Nokia will provide data-centre networking alongside Bull’s BXI high-performance interconnect.

The configuration highlights the nuances behind Europe’s push for technological sovereignty. The system will be designed and integrated by a European supplier, with European networking and interconnect technology playing significant roles. But the accelerators at the heart of its AI performance will still come from US chipmaker AMD, with IBM providing another critical layer of the infrastructure.

Building more computing capacity on European soil therefore reduces one strategic constraint (access to computing power) without eliminating dependence on foreign semiconductor and infrastructure suppliers. Bull also plans to use direct warm-water cooling and recover waste heat from the supercomputer for Kajaani’s district-heating network. The company says the facility will run entirely on renewable electricity, continuing the Finnish site’s emphasis on energy-efficient high-performance computing.

An early test for the newly state-owned Bull

The contract carries particular weight for France. The French government completed its acquisition of Bull from Atos on March 31, in a transaction agreed at an enterprise value of up to €404 million. Bull generated around €720 million in revenue in 2025 and employs more than 3,000 people, including a large concentration of Europe’s supercomputing expertise.

LUMI-AI is therefore one of the first major commercial tests of Bull since the company returned to direct French state ownership. It also comes as EuroHPC’s role expands beyond its original scientific-computing mission toward providing AI infrastructure for companies. Europe is establishing 19 AI Factories and 13 associated “antennas”, with start-ups, SMEs and researchers among the intended users of the network.

The programme is already moving to a larger scale. In July, EuroHPC opened its call for future AI Gigafactories (substantially larger facilities intended to support frontier-model training, fine-tuning and inference). European authorities expect the scheme to help unlock more than €20 billion of private investment, with proposals due by November 12, 2026.

LUMI-AI therefore addresses only part of Europe’s computing gap. The next test will be whether the continent can turn publicly funded machines into computing capacity that is sufficiently large, accessible and rapidly usable for European AI developers competing with companies backed by far larger private infrastructure.

The company’s European pipeline is not limited to AI factories: Spain’s AEMET is also deploying a Bull supercomputing system for operational weather and climate forecasting.

By Hugo Noirtault

Founder of Stack Chronicle and a journalist with experience across several local and national radio stations (including RTL2, Hit West, Euradio and Figaro Radio in France, and WDET in Detroit, Michigan), I began taking a deeper interest in technology in the wake of the COVID-19 pandemic. I quickly realised, that few media outlets were building meaningful connections between technology, AI, finance, politics and environmental issues. That observation led to the creation of Stack Chronicle: a publication designed to help both curious readers and more informed audiences understand the technological transformations shaping our world.

Leave a Reply

Your email address will not be published. Required fields are marked *