At the GTC 2026 conference, NVIDIA officially unveiled "Vera Rubin," its next-generation AI computing platform. This announcement signals a strategic shift toward the "AI Factory" concept, where data centers are no longer viewed as mere collections of servers but as single, massive computing engines.

The core of the system integrates the latest Vera CPUs and Rubin GPUs via the NVLink 6 high-speed interface. By moving from chip-based delivery to integrated rack-level systems, NVIDIA is providing a unified foundation for AI training, inference, and the execution of sophisticated AI agents. Jensen Huang (CEO of NVIDIA) emphasized this transformation, stating that the entire data center has effectively become the smallest unit of computation.

On the software front, the company introduced "NemoClaw," an open-source framework built for "always-on" agents that can perform tasks autonomously without constant human intervention. A key feature of this platform is its flexibility; it allows for hybrid operations where sensitive data is processed locally while complex reasoning is offloaded to the cloud, all manageable via a single command.

Furthermore, NVIDIA is expanding its reach from Earth to orbit with the introduction of "DLSS 5" and the "Space-1 Vera Rubin Module." DLSS 5 represents an evolution into neural rendering, where AI interprets scene semantics to generate lighting and textures from scratch. Meanwhile, the space-specific modules accelerate "On-Orbit AI" capabilities, enabling satellites to process data and make decisions independently of terrestrial infrastructure.

At the GTC 2026 conference, NVIDIA officially unveiled "Vera Rubin," its next-generation AI computing platform. This announcement signals a strategic shift toward the "AI Factory" concept, where data centers are no longer viewed as mere collections of servers but as single, massive computing engines.

The core of the system integrates the latest Vera CPUs and Rubin GPUs via the NVLink 6 high-speed interface. By moving from chip-based delivery to integrated rack-level systems, NVIDIA is providing a unified foundation for AI training, inference, and the execution of sophisticated AI agents. Jensen Huang (CEO of NVIDIA) emphasized this transformation, stating that the entire data center has effectively become the smallest unit of computation.

On the software front, the company introduced "NemoClaw," an open-source framework built for "always-on" agents that can perform tasks autonomously without constant human intervention. A key feature of this platform is its flexibility; it allows for hybrid operations where sensitive data is processed locally while complex reasoning is offloaded to the cloud, all manageable via a single command.

Furthermore, NVIDIA is expanding its reach from Earth to orbit with the introduction of "DLSS 5" and the "Space-1 Vera Rubin Module." DLSS 5 represents an evolution into neural rendering, where AI interprets scene semantics to generate lighting and textures from scratch. Meanwhile, the space-specific modules accelerate "On-Orbit AI" capabilities, enabling satellites to process data and make decisions independently of terrestrial infrastructure.