On 13 October in Paris, Jensen Huang did something tech chiefs rarely do: he announced the total before the cheques. The Nvidia CEO promised $2 billion in European deals within a week, on top of the $8 billion already committed on the continent over the past twelve months. He also named the first beneficiaries: four French startups. Behind the enthusiasm lies a question: is Europe financing its independence, or its chip supplier's order book?
What was announced
Nvidia announced planned minority investments in Doctolib (healthcare), Genesis AI (robotics), Pigment (business planning software) and Entalpic (AI-driven materials discovery). No amounts were disclosed. According to AFP, Huang wants these stakes to "kickstart the artificial intelligence revolution in Europe". In Paris he also said: "Europe needs to get in the race", Quartz reports.
Nvidia now puts its European commitment at nearly $10 billion, without separating equity stakes from infrastructure. The same day, it announced in the UK a compute donation worth $250 million over five years for research on AI agent safety and security.
Doctolib: the telling example
In its press release, Doctolib says the investment will fund AI hiring and the training of its models in Europe with help from Nvidia's experts. The company, which says 100 million patients and 580,000 health professionals use its services, has used Nvidia technology for a year. It plans to build models based on open models (open-weight and open-source) and to host all health data in the European Union and the UK, under its sole control.
"We are convinced that AI can have a tremendous impact for patients and health professionals."
That sentence from Stanislas Niox-Chateau, Doctolib's co-founder and CEO, sums up the promise: European data and European models. But as Tech Funding News points out, the hardware remains American: the data and the models stay European, the chips do not.
Sovereignty through a single supplier
Euronews notes that France is betting on partnerships rather than full autonomy: Foxconn and Bull assemble AI machines in Angers using Nvidia chips, and Bull is to build systems based on the Vera Rubin Ultra NVL72 platform there from 2028. Nvidia also showcased its open Nemotron models, already used by French organisations, and the Twake.ai workplace suite (Linagora and Docaposte), recommended as the "sovereign" digital workplace for more than 2,500 public bodies.
That is where the debate lies. According to Counterpoint Research, cited by Tech Funding News, 92% of "sovereign" large language models across more than 80 countries outside the US and China run on Nvidia chips. And every startup Nvidia funds is also a customer for its chips. A Forrester analyst, cited by Clubic via Tech Funding News, argues that such stakes tie partners' tools to Nvidia's architecture and make a later switch to AMD costlier.
The real problem: the capital gap
The underlying issue is money. According to the OECD, AI companies took $258.7 billion of global venture capital in 2025, or 61% of the total. US-based firms received about $194 billion (75%), versus $15.8 billion (6%) for EU-based firms: roughly twelve times less. Against such a gap, an investor ready to put billions into Europe is hard to refuse.
Who wins, who loses
- Nvidia wins: the image of a partner of Europe, loyal customers and access to top teams.
- The four startups win: capital, technical expertise and credibility with other investors.
- European sovereignty is mixed: data and models stay local, but the dependence on the same compute supplier remains.
- Alternatives (other chipmakers, European compute players) may struggle more if ecosystems tune themselves to Nvidia.
What it means for us, students
Doctolib, Pigment, Entalpic and Genesis AI are hiring AI talent: a signal for those torn between leaving for the US and staying in Europe. It is also a textbook case for public policy, law or digital economics programmes: you can have European data and models and still depend on a foreign supplier. Two things to watch in the coming weeks: the real size of Nvidia's cheques, and whether these startups can build on its chips without giving up most of their value.
Sources: Doctolib, press release, 13 October 2026; Tech Funding News, 14 October 2026; Euronews, 14 October 2026; Tech.eu, 14 October 2026; Quartz, 13 October 2026; AFP via Boursedirect, 14 October 2026; NVIDIA Blog. Photo: Anderseidesvik, Wikimedia Commons, CC BY-SA 4.0.
