Nvidia has unveiled its smallest AI supercomputer to date, the NVIDIA DGX Spark. The device is powered by the GB10 Grace Blackwell Superchip, comes preinstalled with Nvidia’s AI software stack, and offers 128GB of memory. Elon Musk is the recipient of the very first unit, which was hand-delivered to him by Nvidia CEO Jensen Huang yesterday at the SpaceX Starbase in Texas.

The GPU and AI giant detailed the delivery process in a recent blog post; “Imagine delivering the smallest supercomputer next to the biggest rocket,” remarked Huang with a chuckle. The Nvidia CEO also stopped by to greet staff, kids, and grab a slice of pizza from the cafeteria.

Nvidia DGX Spark is a compact AI supercomputer

DGX Spark is the latest of Nvidia’s AI hardware, coming nine years after the DGX-1, which marked a commitment to AI long before its rivals – and it has paid off big time. The new AI supercomputer was also signed by Huang.

• NVIDIA GB10 Grace Blackwell Superchip — delivering up to 1 petaflop of AI performance at FP4 precision.
• 128GB of unified CPU-GPU memory — so developers can prototype, fine-tune and run inference locally without bouncing between machines or cloud instances.
• NVIDIA ConnectX networking for clustering and NVIDIA NVLink-C2C for 5x PCIe bandwidth.
• NVMe storage for speed and HDMI out for visuals.

Source: Nvidia

Designed for developers, researchers, and creators, the DGX Spark offers “supercomputer-class performance” on the go. The special delivery came in time for SpaceX’s Starship Flight 11 test launch yesterday (October 13) at 6:15 PM CDT. Following the launch, SpaceX says the 11th flight test “reached every objective“.

Its compact size – the size of “a piece of origami paper, and the thickness of a hardcover book” is a milestone for this kind of computing power. The AI supercomputer comes housed in a 1.2kg chassis. The Nvidia AI software stack unlocks access to frameworks, libraries, pre-trained AI models, and Nvidia’s NIM microservices.

About the Author

Jack Goodall

At PC Guide, Jack is mostly responsible for reporting on hardware deals. He also specializes in monitors, TVs, and headsets and can be found putting his findings together in a review or best-of guide.