LectureMaster
Has Man Musk

NVIDIA’s Jensen Huang Hand-Delivers the “World’s Smallest Supercomputer,” the DGX Spark, to Elon Musk, Right Alongside Its Retail Launch
NVIDIA's CEO has delivered the DGX Spark to billionaire Elon Musk, highlighting the importance of the device in Jensen's eyes.


NVIDIA's CEO has delivered one of the firm's most optimistic products, the DGX Spark, to billionaire Elon Musk, highlighting the importance of the device in Jensen's eyes.
NVIDIA's DGX Spark To Launch Into Retail By October 15th, Bringing Immense AI Power to Users
The DGX Spark is known to be one of the most compact devices available, offering computational performance that no one could have ever thought possible. The product aligns with Jensen's core idea of making AI accessible to everyone. The DGX Spark was showcased at CES 2025, and now, NVIDIA's CEO has managed to deliver a unit of the mini-supercomputer to Elon Musk, right around the 11th test of Starship. This move also took us on a trip down memory lane, to a time when NVIDIA's CEO delivered one of the first units of the DGX-1 to Musk during his tenure with OpenAI.Huang arrived walking past rows of engineers who waved and grinned. Moments later, Musk appeared in the cafeteria, greeting staff and opening donuts and chips for kids before grabbing a slice of pizza.
Huang joined him, recounting the story of delivering the first DGX system to OpenAI and explaining how Spark takes that mission further.
- NVIDIA
NVIDIA's DGX Spark is now expected to be available through retailers by October 15th. It can be ordered directly from the firm's official website, along with vendor availability from Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI. We have deep-dived into relevant SKUs and the specifics of each custom model in a previous post here, but for a quick rundown of the DGX Spark specifications, here are the main highlights:
- 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.

Image Credits: NVIDIA
The mini-supercomputer experienced a slight delay in its launch, as retail availability was previously expected in July. However, given the complexities involved with the custom NVIDIA and MediaTek GB10 SoC, the launch was pushed ahead to October. The device offers immense performance for professionals, bringing massive AI computational power in the palm of your hands, but it also features a hefty price tag, right around $3,999.