NVIDIA DGX Station with GB300 Superchip Arrives at Developers' Desks

At GTC 2026, NVIDIA announced that the first DGX Station systems – powered by the NVIDIA GB300 superchip – have now reached pioneering AI developers. The milestone marks a significant step forward in deskside development with frontier-scale models.

DGX Station is NVIDIA's most powerful workstation-format AI system: essentially a server in a desktop form factor, designed for researchers and engineers who need local access to frontier models without relying on cloud infrastructure. The GB300 chip is the successor to Blackwell, representing a major leap in compute density and efficiency.

What this means in practice

With the GB300 DGX Station, large models can be run locally at speeds and memory capacities that previously required data center configurations. For organizations that want to keep sensitive data on-premise – such as companies with ISO 27001 compliance requirements or GDPR data residency needs – this is a compelling alternative to cloud-based inference.

NVIDIA has positioned the DGX Station as the tool for those building "the next generation of AI," and its arrival at developers' hands during GTC signals the hardware is ready for production.

GTC 2026 is ongoing in San Jose, with more announcements expected throughout the week.