NVIDIA's European 'AI Masterplan' Includes 20 Factories, Quantum Leaps & 10X Computing Power, According To Jensen Huang

On Wednesday, NVIDIA CEO Jensen Huang delivered a GTC keynote speech in Paris, France. Beyond quantum computing remarks that stirred market movements, the computing chip giant also announced a series of European AI infrastructure partnerships, signaling its ambitious expansion across the continent.
Quantum Computing: A Turning Point?
Huang dedicated several minutes to promoting NVIDIA's hybrid quantum-classical computing platform, CUDA-Q, announcing that the company's Grace Blackwell systems now support this open-source quantum development platform. He stated, "Quantum computing is reaching an inflection point."
He added that humanity is on the verge of "harnessing quantum computers that use qubits as information units," and these technologies will "be applied across multiple fields and solve some fascinating problems in the coming years." In next-gen supercomputers, QPU (Quantum Processing Units) will work alongside GPUs-QPUs handle quantum calculations while GPUs manage preprocessing, control, error correction, and post-processing.
These remarks triggered pre-market rallies for quantum computing stocks including IonQ (+4.93%), Quantum Computing Inc.( +5.51%), and Rigetti Computing (+6.75%).
Earlier this week, NVIDIA announced collaborations with IonQ, AWS, and AstraZeneca to develop quantum-accelerated computing workflows for chemistry applications.
This comes after Huang's January statement that "practical quantum computers are still 20 years away." He later apologized at March's GTC conference while unveiling a new quantum computing research lab in Boston, partnering with Harvard and MIT scientists.
Europe's AI Factory Boom
Huang's European tour primarily highlighted NVIDIA's growing partnerships in the region. He noted that Europe has recognized the importance of "AI factories" (NVIDIA's term for large-scale data centers), with 20 major AI factories currently under construction across the continent. He projected that Europe's AI computing power will increase tenfold within two years.
Huang also announced a collaboration with French startup Mistral AI to build an end-to-end cloud platform. Follow-up disclosures revealed that the first phase of the partnership will deploy 18,000 sets of NVIDIA's Grace Blackwell systems, with plans to expand to more locations by 2026.
NVIDIA is also working with UK cloud service providers Nebius and Nscale to construct AI infrastructure. The two companies had previously announced plans to deploy 14,000 Blackwell GPUs.
As one of Wednesday's key announcements, Jensen Huang revealed that NVIDIA will build the world's first industrial artificial intelligence cloud platform to help European industrial enterprises accelerate full-process manufacturing applications - from design and engineering simulation to digital factory twins and robotics technology.
The project will be located in Germany and powered by DGX B200 systems equipped with 10,000 Blackwell GPUs, along with RTX PRO computing support.
Sovereign AI & European Tech Hubs
NVIDIA is also working with telecom giants in France, Italy, Norway, Switzerland, and Spain to deploy AI infrastructure and develop sovereign AI models-ensuring data centers and servers remain within Europe. The company is setting up tech hubs across the region to accelerate breakthroughs in the UK, France, Spain, and Germany.
Additionally, NVIDIA launched the Nemotron toolkit to help European developers build region-specific large language models (LLMs). These models will integrate with Perplexity's AI-powered search, allowing users to "ask and receive answers in their native language, cultural context, and thought patterns."
Pharma & Robotics: The Next AI Wave
In a surprise move, NVIDIA announced a partnership with Danish weight-loss drug giant Novo Nordisk to develop custom AI models for early-stage clinical trials.
Huang concluded by declaring that the next wave of AI is already here-and growing exponentially: "We have physical robots, and we have information robots. We call them agents. The technology necessary to teach a robot to manipulate, to simulate - and of course, the manifestation of an incredible robot - is now right in front of us."
Comments
No comments yet