Nvidia chief Jensen Huang delivered a bold prediction during Wednesday’s GTC conference in Berlin: every European company will eventually transform into an AI company. The declaration accompanied a slate of new partnerships aimed at strengthening the continent’s position in robotics and physical AI.
Europe has struggled to keep pace with American and Asian rivals in artificial intelligence. No homegrown firm matches the scale of OpenAI, Anthropic, or Nvidia itself. Huang, however, sees a different kind of advantage. He pointed to Europe’s unmatched precision in mechatronics, electronics, and mechanical engineering as the foundation for leadership in physical AI.
“To be part of this AI economy, you don’t want to just be consuming the AI; you want to be producing the AI,” Huang said. “We could do all of that here in Europe.”
The chipmaker has now committed nearly $10 billion to Europe’s technology ecosystem. Its early dominance in supplying processors and software for robotic “brains” gives Nvidia a direct commercial stake in the region’s industrial success.
Among the newly announced collaborations, Germany’s Agile Robots plans to manufacture between 1,000 and 2,000 compact humanoid robots next year in partnership with Nvidia. Production launches late this year, according to CEO Zhaopeng Chen.
Huang also unveiled a new DSX POD design for factory-built data center modules running on the Vera Rubin platform. Deployment time drops from six months to roughly two weeks, Nvidia said.
Meanwhile, a stage demonstration highlighted both European creativity and the stubborn limits of current robotics. French startup Pollen Robotics, working with Hugging Face, presented tiny duck-shaped AI-trained robots called Microducks. When Huang instructed them to nap, three toppled over immediately. A fourth refused, requiring repeated commands.
The last robot then faced the wrong direction as Huang finished his remarks. “Pay attention, stay focused,” he told it sternly.
The moment underscored a reality beneath the sweeping vision: Europe may hold engineering talent, but teaching machines to behave reliably remains a work in progress.
