Mistral is also collaborating with Bpifrance, MGX and NVIDIA on the expansion of Campus AI, an initiative centered on a planned 1.4-gigawatt facility that would rank among Europe’s largest AI campuses.
The projects are part of a wider buildout of AI infrastructure across France. Scaleway has introduced NVIDIA Blackwell B300-SXM instances through its cloud platform, while Bull and Foxconn have unveiled plans to manufacture NVIDIA Vera Rubin NVL72 systems in Europe. Under the arrangement, systems will be produced and initially tested in the Czech Republic before final assembly and validation in France.
France is also positioning itself as a candidate location for future large-scale AI infrastructure. A consortium of eight French companies has submitted a proposal to host a European AI gigafactory, while Schneider Electric and NVIDIA are working together on designs intended to support gigawatt-scale AI facilities.
Alongside investments in computing infrastructure, French organizations are advancing open AI models and development platforms designed for regional languages, cultural requirements and regulatory frameworks.
At VivaTech, executives from Gradium, H Company, LINAGORA, Pleias and NVIDIA discussed how open models can support greater transparency, customization and local relevance for governments, businesses and developers.
“What we see now is a shift from building one isolated model to running continuous model infrastructure, where models train the next models, curate data, generate synthetic environments and verify reinforcement learning,” said Pierre-Carl Langlais, chief technology officer of Pleias. “Open model infrastructure is simply the way to ensure that many people can build AI and frontier-level practice can disseminate throughout the entire economy.”
Several French AI companies are contributing to that effort through the NVIDIA Nemotron ecosystem. Mistral, a founding member of the NVIDIA Nemotron Coalition, is contributing expertise in model development and multimodal systems. LINAGORA is developing multilingual large language models through its Luciole family, with a particular emphasis on French-language capabilities and local cultural context. The company’s Luciole 1B, 8B and 23B models were pretrained on the Jean Zay supercomputer through the OpenLLM-France project and are available as open-source releases alongside their training datasets.
H Company is developing Holotron, a family of AI agents built on NVIDIA Nemotron models that can interact with software interfaces without requiring APIs or custom integrations. Pleias, meanwhile, has worked with NVIDIA to create synthetic persona datasets based on French and Belgian demographics and cultural context. The company is also training compact language models on open datasets and developing specialized systems for search, retrieval-augmented generation and public-sector document workflows.
The expansion of infrastructure and model development is increasingly being matched by production deployments.
AI Factory France, led by GENCI, has partnered with NVIDIA Inception and NVIDIA Connect to provide startups with access to national supercomputing resources, including Jean Zay. Early participants such as Pleias, Nebula and Ryax Technologies are already using that access to develop deployable applications.
Across industry, companies are integrating AI into operational workflows. Sanofi is deploying AI agents across research, manufacturing, commercial operations, procurement and IT functions, while also working with Owkin and Biolevate on autonomous systems for drug discovery and development.
Orange Business has expanded internal use of its Live Intelligence GenAI platform, reporting more than 100,000 active users within the company. The platform is also being offered to businesses and public-sector organizations across Europe, with data hosted within the region.
In manufacturing, Stellantis has launched an initiative focused on AI-powered digital twins across its operations. Dassault Systèmes is combining virtual twins, AI infrastructure and open models through its 3DEXPERIENCE platform to support the design, simulation and operation of complex systems.
Elsewhere, TotalEnergies is building the Pangea 5 supercomputer with Dell Technologies and NVIDIA to support seismic imaging, simulation and AI-driven research. L’Oréal is using its CreAltech platform to combine generative AI with 3D digital twins to help creative teams expand content production while maintaining consistency and responsible AI practices.
Taken together, the developments highlight a broader transition underway in France’s AI sector. Infrastructure projects that were previously announced are entering operation, open-model initiatives are expanding, and enterprises are moving AI systems from pilot programs into production environments across a range of industries.
This analysis is based on reporting from Nvidia.
Image courtesy of Unsplash.
This article was generated with AI assistance and reviewed for accuracy and quality.
