NVIDIA to Announce $2 Billion in New European AI Investment, on Top of $8 Billion Deployed Over the Past Year

NVIDIA hints at announcing $2 billion in new European AI investments, adding to $8 billion already deployed, as Jensen Huang keynotes at NVIDIA GTC Berlin next week.

Earlier this week, NVIDIA CEO Jensen Huang signaled that the company might announce $2 billion investment in new European AI deals at the NVIDIA GTC event in Berlin next week. That comes on top of roughly $8 billion the company has deployed across the continent over the past year, which brings NVIDIA’s stated European commitment to nearly $10 billion.

Europe needs to get in the race,

says Jensen in Paris, the first stop of Europe’s tour that continues to London and Berlin, where he keynotes GTC Berlin from 20 to 22 October.

European startup investments announced

The Paris announcements named four French startups for planned minority investments, each in a different corner of the economy:

Doctolib builds agentic tools for healthcare professionals and a digital companion for patients.
Entalpic applies AI to atomic-scale materials discovery, with a focus on semiconductor manufacturing.
Genesis AI is developing general-purpose robots on its own foundation models.
Pigment provides AI-driven planning software for finance, sales and operations.

Across the Channel, NVIDIA named Boltz, Foresight, Hiverge, Prima Mente and Riverlane. In the Nordics it announced a Stockholm AI Lab with the Norrsken Foundation, due to open in 2027, and a collaboration with Sweden’s EuroHPC-funded AI Factory to open infrastructure to startups and SMEs.

Why this isn’t just venture capital

NVIDIA is building at every layer of the stack: capital for founders, GPUs for builders, research partnerships, developer communities and open models for everyone. Its regional cloud partners, including Nscale, Verda and Volta, are working toward roughly one gigawatt of AI capacity in the Nordics by the end of 2027.

The pitch is sovereignty.

Open models are important as they can be customised for languages and for your own sovereignty,

he said.

Mistral AI, NVIDIA’s French open-weight partner, is the proof of concept. Its latest model was trained on 4,000 Grace Blackwell chips.

Europe is one of the great regions of industrial manufacturing,

Jensen said at the Paris reception, adding that

this is the beginning of the AI journey for the EU.

The physical AI thesis

Deepu Talla, NVIDIA’s Vice President of Robotics and Edge Computing says, “Physical AI is the next big opportunity.”

His reasoning is simple: around 70% of GDP sits in the physical world. For a region built on factories, ports, vehicles, materials and machinery, that is not an abstract claim. It is a description of where Europe’s strengths already are. Deepu argued that Europe can pair AI with its human capital, engineering expertise and industrial capabilities to create more value.

You can already see the thesis in practice. In France’s Côte-d’Or, a digital-twin agent built on NVIDIA’s Cosmos and Nemotron models manages nearly 5,800 kilometers of roads and buildings and has cut operating costs by 10%.

Dassault Systèmes is training its own industrial world models on NVIDIA technology. Loft Orbital’s Altair-1 satellite is running Mistral’s models on NVIDIA Jetson modules in orbit.

The tradeoff

The same sovereignty narrative sits next to a plain fact: the more European startups build on NVIDIA’s silicon, software and capital, the more they depend on a single US supplier. The open question is how much of the money circles back into NVIDIA hardware orders.

There is also a reality check inside the portfolio. Doctolib, a headline name in the French cohort, was valued at €3.6 billion in a secondary sale this year, well below the €5.8 billion it commanded in 2022. Backing a company is not the same as guaranteeing its trajectory. And public skepticism about AI is rising across Europe, driven by concerns over autonomous agents and the energy demands of data centers.

Huang has resisted calls to slow down, arguing that the technology will strengthen economies and can be managed responsibly.

What founders and policymakers should do

Founders should take the capital, the compute and the credibility, but build for portability and treat open models as an insurance policy, not just a cost saver. Policymakers should see this week as leverage. The EU’s AI Gigafactories program and national compute strategies are where Europe decides whether NVIDIA’s investment becomes a foundation or a dependency.

Jensen has said NVIDIA expects to sell twice as many chips next year as this year. Europe is clearly part of that plan. The question for the continent’s founders and governments is whether they are part of the plan on their own terms.