At NVIDIA GTC Paris 2025, NVIDIA CEO Jensen Huang declared that Europe isn’t just adopting AI—it’s building it. Speaking from the Dôme de Paris during VivaTech, he called artificial intelligence the new “intelligence infrastructure,” shaping the future of nations and industries.
Central to this transformation is the GB200 NVL72, a powerful AI system Huang described as “one giant GPU.” Built to reason and plan, it now supports sovereign models and quantum computing. NVIDIA’s partners currently produce 1,000 GB200 units weekly, showcasing rapid scaling. Alongside it, the smaller DGX Spark and RTX PRO Servers serve various AI deployment needs, from startups to national infrastructure.
Huang said countries need two approaches: AI infrastructure for public innovation and AI factories for proprietary operations. Both rely on NVIDIA platforms, and governments across Europe are now deploying them. NVIDIA is also expanding its tech center network, with new hubs opening in Finland, Germany, Spain, Italy, and the UK to accelerate local talent and drive quantum innovation.
Europe’s quantum journey advanced further with the launch of CUDA-Q on Denmark’s Gefion supercomputer. Huang also announced CUDA-Q’s availability on Grace Blackwell systems, enabling hybrid classical-quantum applications. These partnerships with supercomputing centers will fuel error correction research and practical use cases.
To meet regional AI needs, Huang introduced NVIDIA Nemotron, which allows developers to create multilingual large language models suited to local culture. These models are being used in Perplexity, a European search engine, enabling localized and private AI deployments.
For companies building agents, Huang unveiled new tools including the NeMo Agent toolkit and an AI Safety blueprint. These resources help organizations design smarter, safer agents powered by data flywheels and reinforced learning loops.
NVIDIA is now rolling out the DGX Cloud Lepton platform across Europe. This enables instant access to accelerated compute and seamless deployment. Integrated with Hugging Face, it allows developers to run models anywhere using a single architecture and deployment process.
In summary, NVIDIA GTC Paris 2025 demonstrated how Europe is shaping the AI frontier. With high-powered GPUs, quantum-classical platforms, and region-specific AI models, Europe is not just participating in the AI revolution—it is leading it.
READ: AI Art Gallery at GTC Paris Redefines Creative Expression
