SoftBank Group plans to invest up to €75 billion to build 5 gigawatts of AI data center capacity in France, positioning the country as a major European hub for AI infrastructure.
SoftBank's massive investment underscores France's emerging role as a critical player in Europe's AI infrastructure race. The Japanese conglomerate will develop and operate up to 5 gigawatts of additional data center capacity, a significant expansion of computational resources needed to power large language models and AI applications.
The investment reflects broader competitive dynamics in AI infrastructure development. Tech giants and infrastructure investors are racing to secure capacity for training and deploying increasingly resource-intensive AI models. Data centers require substantial power, cooling systems, and real estate—making location and government support crucial factors.
France offers strategic advantages for such development. The country has committed to supporting AI innovation and possesses relatively abundant hydroelectric power in certain regions, which can reduce operating costs compared to other European locations. Additionally, European regulatory frameworks around AI are shaping investment decisions, with firms seeking to establish infrastructure within the continent to serve European markets.
The investment also comes as major AI companies expand their infrastructure footprints. Salesforce recently deepened its stake in AI startup Anthropic to approximately $5 billion, illustrating continued capital flow into AI ventures. Such investments signal confidence in AI's commercial viability and the infrastructure required to support it.
SoftBank's commitment extends beyond immediate capacity needs. By establishing substantial data center operations in France, the company positions itself to serve European AI development and deployment for years to come. The scale—5 gigawatts—represents meaningful computational power that could support numerous AI applications across industries.
The move also addresses European concerns about AI infrastructure dependency. Rather than relying solely on U.S.-based data centers, European stakeholders can leverage locally operated facilities, potentially reducing latency and addressing data sovereignty considerations.
Investments of this magnitude typically involve partnerships with local governments and power providers. The specific timeline and structure of SoftBank's French project remain subjects for further development as plans advance.
Journalist Joanna Stern spent 2025 integrating artificial intelligence into every aspect of her life—from text responses to meal decisions to book editing—documenting what worked, what didn't, and what surprised her most.
Google has released macOS versions of AI Edge Gallery and AI Edge Eloquent, enabling users to run open AI models locally on their devices. The tools include support for Google DeepMind's Gemma 4 12B model.
Anthropic has published technical details on how it isolates Claude AI across different product deployments, addressing safety and security concerns in multi-product environments.
Leading mathematicians are raising concerns about the rapid pace of AI development and its potential implications for the field. The warning comes as artificial intelligence systems continue to expand capabilities across mathematical problem-solving domains.