Nvidia has quietly taken minority positions in nine European startups across France and the UK, declining to reveal how much money changed hands or what any of the companies are now worth.
The chipmaker’s move lands in the middle of a European AI funding surge. Mistral AI just closed a €3 billion Series D, Oxford photonics startup Salience Labs pulled in $125 million, and defence AI player Helsing raised €1.8 billion at an €18 billion valuation.
The French investments span four companies. Genesis AI builds robotic foundation models alongside a physics simulation engine that generates training data. It emerged from stealth in July 2025 with a $105 million seed round co-led by Eclipse Ventures and Khosla Ventures. Bloomberg has since reported talks for a $500 million round at a $3 billion pre-money valuation, though the company has not confirmed those discussions.
Pigment, founded in 2019, sells planning software for finance, sales and operations teams. Its $145 million Series D in April 2024 pushed it into unicorn territory. Doctolib, the largest of the French group by user reach, counts 100 million patients and 580,000 healthcare professionals on its platform. The company logged €422 million in annual recurring revenue for 2025 and says it operates profitably. Meanwhile, Entalpic applies AI to materials discovery for semiconductors and electronics, a compute-heavy task that aligns closely with Nvidia’s core business.
On the UK side, Boltz develops open-source models that predict how proteins fold and bind. Its $28 million seed round arrived in January 2026 at a reported $125 million valuation, the same day it announced a Pfizer collaboration. Riverlane tackles quantum error correction, the layer that prevents fragile qubits from corrupting calculations. Nvidia’s own blog notes that GPUs have accelerated the recalibration of Riverlane’s error-correction decoders.
Foresight forecasts delivery dates for massive construction projects including data centres, claiming twice the accuracy of conventional methods. Hiverge, a Cambridge algorithm lab, reports a 15% gain on AI training benchmarks and a 10,000x speed-up on aircraft-loading problems. Faster algorithms mean fewer GPU hours per result, an interesting tension for a chip seller. Prima Mente builds foundation models on epigenetic data, with its Pleiades model scoring 0.89 accuracy on early Alzheimer’s and Parkinson’s detection.
Nvidia has also pledged $250 million in computing resources over five years to support UK AI safety research. The investments collectively signal the company’s deepening bet on European AI infrastructure and application layers, even as individual deal terms stay hidden.
