French orgs operationalize sovereign AI via Bull servers, Altair-1 satellite and NVIDIA digital tools

Organizations across France are deploying agentic computing systems in space, finance and public services, marking a shift from strategic planning to operational reality. According to NVIDIA blog research, this ecosystem expansion relies on deliberate groundwork, including compute commitments anchored by the AI Action Summit in February 2025 and subsequent Choose France Summits.

The deployment spans the full technology stack, with companies building infrastructure on NVIDIA platforms. This includes hardware manufacturing, satellite operations and software services designed for sovereign data requirements. The trend demonstrates how French enterprises are integrating advanced modeling and autonomous agents into core business functions and public sector operations.

What we know

Hardware production is scaling within Europe, with Bull extending its AI server manufacturing to include NVIDIA Vera Rubin Ultra NVL72 systems. These systems are planned for release in 2028, with tray production in Europe and rack integration occurring at the Angers factory, which was inaugurated on October 2.

In the space sector, Altair-1, the first satellite in a $1 billion constellation, launched this month and reached orbit. Built by Orbitworks, a joint venture between Loft Orbital and Marlan Space, the satellite uses NVIDIA Jetson AGX Orin modules running Mistral models to perform real-time Earth observation and autonomous orbital operations.

Public sector adoption is significant, with LINAGORA and Docaposte securing the CANUT framework agreement. This makes Twake.ai, built on NVIDIA Nemotron open models, the sole referenced sovereign digital workplace for more than 2,500 French public organizations, ensuring trusted data handling for government entities.

Financial services are also participating, with SBS launching a sovereign platform for banks to deploy governed generative and agentic services. This collaboration with NVIDIA allows financial institutions to run these services within their own secure environments, maintaining control over sensitive data while leveraging advanced AI capabilities.

Retail and telecommunications sectors are seeing measurable efficiency gains. L’Oréal uses the NVIDIA ALCHEMI Toolkit for in-silico screening, targeting discovery speeds up to 100 times faster. Meanwhile, Bouygues Telecom launched AI Studio V2, powered by NVIDIA infrastructure, enabling employees to build their own agents securely.

Other industries are integrating these tools for specific operational benefits. Mirakl cut seller catalog onboarding to under two hours using fine-tuned models on NVIDIA GPUs. ThinkDeep and Akila built a digital-twin agent managing nearly 5,800 kilometers of roads, reducing operating costs by 10 percent.

Defender's checklist

• Verify that AI agents operate within defined, governed environments to prevent unauthorized actions.
• Ensure that sovereign data requirements are met by using referenced and secure digital workplace platforms.
• Monitor the integration of open models to maintain control over model customization and data privacy.
• Assess the security implications of deploying autonomous systems in critical infrastructure and public services.

Background: Data privacy

Data privacy principles are the rules governing how personal information is collected, used, and stored. They require organisations to limit data collection to what is strictly necessary, keep it secure, and allow users to control their information. Understanding these principles helps you identify when a service respects your boundaries and when it exploits them for profit or surveillance.

Read the full guide: Data Privacy Principles: How to Protect Your Digital Life

What to do and how to stay safe: NVIDIA

• Review current AI deployments to ensure they run within controlled and governed environments.
• Confirm that any sovereign or sensitive data processing uses verified and secure platform references.
• Monitor vendor updates for new security features in agentic computing and open model frameworks.
• Evaluate the operational risks of autonomous systems in critical infrastructure before full deployment.

Step-by-step guide: AI Governance for Small Teams: Practical Controls and Cost

General security guidance from the TechWire Beat newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Which French companies are leading the deployment of agentic AI?

Key players include Bull in hardware, Loft Orbital and Orbitworks in space, and LINAGORA and Docaposte in the public sector, alongside major firms like L’Oréal and Bouygues Telecom.

What infrastructure is powering these agentic AI deployments?

The deployments are largely powered by NVIDIA technology, including Vera Rubin Ultra systems, Jetson AGX Orin modules, and open models like Nemotron and Mistral Large 4.

How many public organizations are using the new sovereign digital workplace?

More than 2,500 French public organizations have secured access to Twake.ai, the sole referenced sovereign digital workplace built on NVIDIA Nemotron.