Summary
NVIDIA says French-linked organisations are applying its hardware and models to satellite operations, public services, retail, research and robotics. The projects range from systems already in use to platforms being launched and hardware planned for 2028.
NVIDIA's October 12, 2026 roundup describes French-linked organisations using its AI hardware and software in satellite operations, public services, retail, telecommunications, research and robotics. The examples are at different stages: a satellite has reached orbit, some services are launching or in production, and one server system is planned for 2028.
The announcement groups a varied set of technologies under agentic computing. Some examples involve agents designed to carry out tasks; others concern model hosting, AI inference, robotics or simulations. They show how organisations are incorporating AI into specific workflows, rather than one common system being deployed across every sector.
From satellite computing to European hardware
Altair-1, described by NVIDIA as the first satellite in a $1 billion, 50-satellite constellation, launched in October and reached orbit. It carries NVIDIA Jetson AGX Orin edge modules running Mistral models for real-time Earth observation and autonomous operations. The satellite was integrated by Loft Orbital and built in Abu Dhabi by Orbitworks, a joint venture between Loft Orbital and Marlan Space.
Edge computing places processing close to where data is collected, rather than relying solely on a distant data centre. On a satellite, that can support analysis and operations in orbit. NVIDIA's announcement identifies the hardware and intended uses, but does not give a performance measurement for Altair-1's onboard AI.
In Europe, Bull is extending its AI server production to NVIDIA Vera Rubin Ultra NVL72 systems, planned for 2028. The company plans to produce trays in Europe and integrate racks at its Angers factory, inaugurated on October 2. The timeline matters: this is a manufacturing and infrastructure plan, not a description of systems already operating at scale.
AI enters public services, retail and research
NVIDIA says its Nemotron open models are in production across French organisations. LINAGORA and Docaposte secured the CANUT framework agreement, under which Twake.ai—built on Nemotron—is the sole referenced sovereign digital workplace for more than 2,500 French public organisations. In financial services, SBS is launching a platform intended to let banks run governed generative and agentic services in their own environments. Mistral Large 4, meanwhile, was introduced under private preview.
The roundup also describes AI being added to existing research and business processes. L'Oréal is integrating NVIDIA ALCHEMI into research workflows so scientists can explore formulation possibilities computationally and use predictions to guide laboratory work. NVIDIA says combining those predictions with experiments can make the discovery process up to 100 times faster than traditional methods.
NVIDIA reports that Mirakl has onboarded more than 70 million products using fine-tuned and open models on NVIDIA GPUs and inference software, reducing seller-catalogue onboarding to under two hours. In Côte-d'Or, ThinkDeep and Akila built a digital-twin agent covering nearly 5,800 kilometres of roads and buildings. NVIDIA reports a 10% reduction in total operating costs for the project, which targets 30% as deployment progresses.
Other examples include Bouygues Telecom and Prisme.ai launching AI Studio V2, which gives employees a controlled environment to build agents, and Paris-based Enchanted Tools deploying social robots to support patients, including in paediatric cancer care. Across these projects, the practical emphasis is integration: placing models and computing within existing scientific, public-sector and operational workflows. The reported gains and capabilities vary by project, and the roundup presents them as company and partner developments rather than results from a single comparative evaluation.