Mistral AI plans 1 gigawatt European computing by 2030, seeks customers.

Mistral AI announced on Tuesday its ambitious plan to build 1 gigawatt of computing power in Europe by 2030. The French artificial intelligence firm aims to solidify European AI sovereignty by offering customers a robust infrastructure with service-level agreements. This expansion includes regional inference endpoints that allow customers to choose whether their AI workloads are processed in Europe or the United States.

The move comes as demand for AI computing resources surges, driven by a growing number of businesses looking to leverage artificial intelligence for various applications. Mistral's expansion strategy features a "Priority Tier" that guarantees uptime for critical deployments. Additionally, the company is forming a coalition of European enterprises, encouraging them to commit to multi-year compute agreements. Mistral expects these commitments to fund 200 megawatts of infrastructure by the end of 2027, laying the groundwork for its ultimate goal of reaching 1 gigawatt by 2030.

In a notable shift, Mistral also plans to host third-party open models on its platform, starting with GLM-5.2 from Z.ai, a Chinese AI lab previously known as Zhipu. This decision may spark debate among those prioritizing strict adherence to European sovereignty, but Mistral argues it will enhance their offerings and attract a broader customer base.

Looking ahead, Mistral's efforts to expand its infrastructure could reshape the landscape of AI computing in Europe. By locking in customers now, the company aims to position itself as a leader in the European AI market, catering to businesses that prioritize both performance and compliance with regional regulations. This initiative reflects a growing trend in the tech industry to balance local data processing needs with global partnerships.