NVIDIA powers French sovereign AI in space, servers and public sector workplaces
Companies in France are moving from ambition to application, using NVIDIA infrastructure and open models for sovereign, secure operations.
Key points
- Bull expands European server production for NVIDIA systems, with integration at its Angers factory.
- Altair-1 satellite launches with on-board autonomous capabilities, running Mistral models on NVIDIA modules.
- Twake.ai becomes the referenced sovereign digital workplace for more than 2,500 French public organizations.
French companies are deploying agentic computing across space, financial services, the public sector and scientific research, according to NVIDIA. The ecosystem has moved from ambition to application, building infrastructure on deliberate groundwork established by recent summits. This shift demonstrates how local firms are scaling autonomous systems for sovereign and commercial use.
Root cause
The progress relies on major compute commitments anchored by the AI Action Summit in February 2025 and the Choose France Summits in 2025 and 2026. These events secured the foundation for a full-stack ecosystem. France-based companies are now expanding agentic infrastructure using NVIDIA technology, enabling real-time operations and sovereign data handling across diverse industries.
Attack path
This section is not applicable as the source material describes infrastructure deployment rather than a security incident. However, the focus on sovereign platforms highlights the need for controlled environments. SBS is launching a platform for banks to deploy governed services within their own environments. Bouygues Telecom allows employees to build agents securely in a controlled setting.
Affected versions
- Bull plans NVIDIA Vera Rubin Ultra NVL72 systems for 2028, with tray production in Europe.
- Altair-1 runs NVIDIA Jetson AGX Orin edge modules with Mistral models for Earth observation.
- Twake.ai is built on NVIDIA Nemotron, serving over 2,500 public organizations under the CANUT framework.
Mitigation
- Use sovereign platforms like SBS’s for governed generative services within bank environments.
- Deploy agents in controlled environments, as seen with Bouygues Telecom’s AI Studio V2.
- Leverage open models like Mistral Large 4 for cyber defense tasks with private preview access.
What to do and how to stay safe: NVIDIA
- Verify that AI agents operate within governed, sovereign environments to protect sensitive data.
- Monitor vendor updates for open models used in critical infrastructure and financial services.
- Ensure that autonomous systems, such as those on satellites, have secure edge computing modules.
- Review access controls for internal AI studios to prevent unauthorized agent creation.
Step-by-step guide: Stop Sensitive Data Leaks in AI Chatbots: Priority Controls
General security guidance from the Vector Bulletin newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Which French companies are using agentic AI?
Companies including Bull, Altair-1, LINAGORA, Docaposte, SBS, L’Oréal, Mirakl and Bouygues Telecom are deploying agentic AI across various sectors.
What infrastructure is supporting these deployments?
The deployments are supported by NVIDIA technology, including Vera Rubin systems, Jetson AGX Orin modules, Nemotron models and full-stack infrastructure.
How many public organizations use Twake.ai?
Twake.ai is the referenced sovereign digital workplace for more than 2,500 French public organizations, secured through the CANUT framework agreement.
