Nvidia introduced DSX Pod, a reference design for modular AI factories, at GTC Berlin on October 21, 2026. Jensen Huang, Nvidia's CEO, stated the DSX Pod aims to accelerate the deployment of AI computing capacity, particularly in Europe, where he noted a current lag in AI infrastructure. The company claims the pre-integrated, containerized design can reduce onsite deployment time for AI data centers from six months to as little as 14 days, provided the site is already prepared with power, cooling, and networking.Nvidia also made Isaac Lab 3.0 generally available, a simulation environment that runs on Omniverse and supports large-scale parallel training. This updated version includes enhanced physics fidelity for manipulation tasks and tighter integration with the GR00T model family. A new teleoperation pipeline within Isaac Lab 3.0 allows operators to collect training demonstrations directly in the simulator. Furthermore, Nvidia introduced OpenShell, an open-source, policy-controlled runtime with kernel-level sandboxing, designed to enhance the security of AI agents by isolating their access privileges. The company also presented the Agent Knowledge Foundation, a set of models and libraries for connecting AI agents to enterprise data, including new Nvidia Kumo models for structured data predictions. Huang emphasized the concept of "sovereign AI," encouraging nations to develop and control their own AI infrastructure, data, and models.
Nvidia Introduces DSX Pod to Accelerate European AI Data Center Deployment
Nvidia CEO Jensen Huang unveiled the DSX Pod at GTC Berlin, a modular data center reference design intended to reduce AI infrastructure deployment times in Europe from months to weeks. The company also expanded its Isaac robotics platform, introducing a new humanoid reference design and updated frameworks for general-purpose robots.