European leaders heard a blunt warning from Nvidia CEO Jensen Huang in Berlin: without a massive computing buildout, the continent risks falling permanently behind in the global AI race. Speaking at the company’s developer and investor conference, Huang identified scarce processing capacity as the single biggest barrier to European competitiveness.
The numbers paint a stark picture. Investment data cited by the Associated Press shows Europe controls roughly 5% of worldwide AI computing power. The United States, by comparison, commands 75%. Janet Tso-Furstenberg, a principal at General Catalyst, reinforced Huang’s assessment, calling limited infrastructure the chief constraint on sector growth across the continent.
Huang’s proposed fix comes in the form of standardized hardware. He introduced the Vera Rubin DSK Pod, a modular computing hub built for rapid deployment by manufacturing partners. These container-sized units ship to prepped locations and can slash data center setup timelines from months down to days. The design targets developers who need to scale operations quickly as demand for large-scale computing accelerates.
Meanwhile, the European AI economy continues pivoting toward industrial use cases, robotics, and physical systems. Nvidia has poured approximately $10 billion into European startups and partners over the past year. The company has also deepened ties with manufacturing heavyweights including Siemens, Mercedes-Benz, and Audi to weave AI into engineering and production workflows.
Security concerns received attention too. Nvidia launched OpenShell, an open-source framework that imposes external constraints on AI application behavior. The system confines those applications within a sandboxed digital environment, limiting unauthorized actions and enforcing security policies.
As Europe grapples with its infrastructure gap, the pressure to act mounts. Huang’s message leaves little ambiguity: build computing capacity now, or watch global AI leadership consolidate elsewhere.
