Nvidia expands AI footprint in Europe with new deals across Europe

Nvidia is accelerating its expansion across Europe, securing a wave of infrastructure and software partnerships to cement its position as a central player in the global AI economy.

Chief Executive Jensen Huang used the company’s GTC event in Paris on Wednesday to reveal a strategy built around “AI factories,” or large-scale GPU-driven data centres, aimed at transforming how European countries build and control their own artificial intelligence capacity.

This marks a deliberate pivot toward what Nvidia calls “sovereign AI,” with regional data centres powering local use cases.

The move comes as Nvidia looks to buffer losses from US export restrictions that limit its high-end chip sales in China.

European data centre deals to boost AI computing

During the keynote, Huang announced that Nvidia expects AI computing capacity in Europe to grow tenfold in the next two years.

The company will achieve this by working with both governments and private sector players across the region.

A major deal involves French AI startup Mistral, which will use 18,000 of Nvidia’s Grace Blackwell chips to deploy a dedicated “AI cloud” for enterprises to train and deploy large language models.

Nvidia is also building a new “industrial cloud” in Germany.

This facility will include 10,000 GPUs specifically tailored for European manufacturers, providing a sovereign computing resource for companies building AI-powered tools and services locally.

Meanwhile, new infrastructure agreements have been signed in Italy and Armenia, underlining Nvidia’s broader push to expand its data centre footprint across both EU and non-EU markets.

Telecom and cloud providers join Nvidia’s AI expansion

To enable regional delivery of AI services, Nvidia is partnering with European telecom firms such as Orange and Telefonica.

These companies will work with Nvidia on the deployment of large language models and AI-based applications across their networks.

The collaboration is designed to speed up the roll-out of new consumer and enterprise-facing services powered by Nvidia’s hardware and software stack.

In addition, Nvidia is integrating its DGX Cloud Lepton product — a global GPU marketplace — with Hugging Face, the AI model repository.

DGX Cloud Lepton allows developers to access GPUs on demand to build and run applications.

The platform is now being expanded through new partnerships with European cloud service providers, which will offer developers regionally hosted compute power for AI workloads.

New research hubs to drive sovereign AI development

Nvidia’s latest announcements also include a wave of new “tech centres” across the continent.

These facilities, to be established in the UK, France, Spain, and Germany, will focus on advanced AI research, training local talent, and supporting innovation ecosystems.

The goal is to foster scientific breakthroughs while enhancing the skills of Europe’s workforce in artificial intelligence, machine learning, and data science.

The broader message behind the company’s European strategy is clear: Nvidia wants to embed itself not just as a chip supplier, but as the backbone of AI infrastructure across the region.

By focusing on localised data sovereignty, Nvidia is offering EU nations a way to maintain control over their AI systems while complying with regional data governance laws.

Software offerings tighten Nvidia’s AI grip

Nvidia is also deepening its software push. Last year, it launched Nvidia NIM — pre-packaged AI models that let developers build applications without designing their own models from scratch.

At the GTC Paris event, Huang confirmed that any large language model available on Hugging Face can now be deployed via NIM.

This integration allows developers to easily access and implement models, further binding Nvidia’s hardware to its expanding software ecosystem.

This software-hardware integration plays a critical role in Nvidia’s strategy to fend off competition and reinforce its dominance in AI development.

By offering full-stack AI tools — from chips to models to deployment frameworks — Nvidia is shaping itself as a one-stop shop for nations and companies seeking to build scalable, sovereign AI.