At Nvidia’s GTC conference in Berlin, CEO Jensen Huang unveiled new hardware and software architectures aimed at accelerating Europe’s AI and robotics ecosystem. Among the announcements was the DSX Pod, a containerized modular data center fitted with Vera Rubin compute racks designed to deploy within 14 days.
Huang emphasized physical AI and robotics as key growth drivers, launching a full-stack robotics platform along with Agile One, a reference humanoid robot scheduled for release in Q1. Nvidia also highlighted autonomous vehicle partnerships with European automakers Mercedes and Volkswagen.
Addressing model ecosystems, Huang predicted that enterprises will increasingly rely on a mix of closed models and customized open-source models tailored to proprietary data. He noted that reasoning AI agents are driving higher compute requirements and sustained GPU demand globally.
Why it matters
Modular, pre-built data center deployments address power grid and planning bottlenecks, shortening time-to-compute for enterprise facilities.
Standardized robotics platforms lower the barrier to entry for physical AI development and humanoid hardware manufacturing.
Enterprise AI strategy is consolidating around hybrid deployments combining foundational closed API models with specialized open-source weights.
Source: wired.com


