Jensen Huang hand-delivers world’s smallest supercomputer to Elon Musk at SpaceX Starbase

Elon Musk, Jensen Huang

The next chapter of the AI revolution just landed in Texas — quite literally. Nvidia CEO Jensen Huang personally delivered the company’s latest innovation, DGX Spark, to Elon Musk at SpaceX’s Starbase facility. The pocket-sized powerhouse is being hailed as the smallest AI supercomputer ever built, capable of running models with up to 200 billion parameters locally.

A supercomputer meets a super rocket

Huang arrived at SpaceX’s gleaming Starbase — flanked by rockets and engineers — to hand over the DGX Spark to Musk. The visit coincided with preparations for the 11th test flight of Starship, the world’s most powerful launch vehicle. Huang reminisced about delivering the first DGX to OpenAI years ago, calling this handoff “a full-circle moment.”

“Imagine delivering the smallest supercomputer next to the biggest rocket,” Huang joked, as he shared pizza and stories with Musk and SpaceX engineers.

What is DGX Spark?

At just 1.2 kg, DGX Spark is roughly the size of a hardcover book but delivers a staggering 1 petaflop of AI performance. Powered by Nvidia’s GB10 Grace Blackwell Superchip, it packs 128GB of unified memory, NVLink-C2C connectivity, and NVMe storage — designed for developers, researchers, and creators who want data-centre-grade performance at their desk.

It ships with Nvidia’s full AI software stack, including Cosmos, Qwen3, and NIM microservices, enabling users to build custom image generators, summarisation agents, or chatbots right out of the box.

A launchpad for the next AI era

DGX Spark isn’t just a machine — it’s a mission statement, according to Nvidia. Nvidia says it represents the shift from cloud-bound AI to local, creative intelligence. In Huang’s words, DGX Spark puts “a petaflop of AI performance within arm’s reach of everyone.”