GPU cloud provider Fluidstack has signed a Memorandum of Understanding with the French government to build an AI supercomputer in the country.

The system was announced at the recent AI Action Summit in Paris and, according to Fluidstack, will be one of the world’s largest low-carbon AI supercomputers.

Founded in 2017 at Oxford University, Fluidstack currently has more than 100,000 GPUs under management on its platform, with customers including Mistral AI, Character.AI, Poolside, and Black Forest Labs.

The facility set to house the system will offer 1GW of dedicated AI compute and will be powered by “France’s abundant, carbon-free, and predominantly nuclear energy,” the company said, adding the site has been designed to scale beyond 1GW by 2028.

No information about the compute hardware has been provided but the system will host “close to 500,000 next-generation AI chips" in its first phase.

The supercomputer is currently set to go live in 2026, with phase one of the project funded by a €10 billion ($10.4bn) investment. Fluidstack said that its financial partners had expressed "strong interest" in funding the project.

The data center will be purpose-built to train advanced AI models and has the potential to create “thousands of high-skilled AI research and infrastructure jobs,” the company added.

“This partnership allows us to rapidly develop the infrastructure needed to support the next wave of AI innovations. Partnering with the French government and global industry leaders, we are creating the most advanced computing capabilities in the world—right here in France,” said Cesar Maklary, co-founder and president of Fluidstack. “France’s commitment to digital and energy sovereignty makes it the ideal location for such an ambitious project, and I am very proud to be playing a key part in this initiative.”

French President, Emmanuel Macron, added: “France is the leading European country in artificial intelligence. Since 2017, we have trained our talents, developed our research, and strengthened our key players in healthcare, space, defense, and large language models. We have a role to play because our nuclear energy is controllable, safe, stable, and decarbonized—ideal for expanding our AI computing capabilities.”

Earlier this week it was reported that Fluidstack had signed a multi-year agreement with French modular data center operator Eclairion to host Mistral’s new training cluster in the Essonne region, south-west of Paris, France.