Texas, October 15, 2025 — NVIDIA Launches DGX Spark has officially begun shipping its revolutionary DGX Spark, the world’s smallest AI supercomputer designed to put petaflop-class AI power right on developers’ desks.
In a headline-making delivery, NVIDIA founder and CEO Jensen Huang personally handed over the first DGX Spark unit to Elon Musk at SpaceX’s Starbase facility in Texas. The gesture paid tribute to history — Huang also hand-delivered the first DGX-1 supercomputer to Musk and OpenAI back in 2016, a move that helped spark the generative AI revolution.
DGX Spark represents a new chapter in personal AI computing. Built on the advanced NVIDIA Grace Blackwell architecture, this compact powerhouse packs up to 1 petaflop of AI performance and 128GB of unified memory in a desktop form factor. With its ability to run inference on models up to 200 billion parameters — and fine-tune models up to 70 billion parameters locally — DGX Spark makes professional-grade AI accessible to individual developers and research teams.
Huang explained the vision behind this next leap in computing innovation:
“In 2016, we built DGX-1 to give AI researchers their own supercomputer,” said Huang. “With DGX Spark, we return to that mission — placing an AI computer in the hands of every developer to ignite the next wave of breakthroughs.”
NVIDIA Launches DGX Spark: A New Era for Desktop AI Computing
DGX Spark brings together NVIDIA’s complete AI platform — including GPUs, CPUs, high-speed networking, CUDA libraries, and the full NVIDIA AI software stack. It transforms a simple desktop setup into a powerful R&D lab capable of developing agentic and physical AI systems.
The system features the new GB10 Grace Blackwell Superchip, NVIDIA ConnectX-7 200 Gb/s networking, and NVLink-C2C technology that offers five times the bandwidth of PCIe Gen5 connections.
Preinstalled NVIDIA AI software allows developers to begin building immediately. Users can explore tools such as NVIDIA NIM microservices, optimize Qwen3 chatbots, experiment with Cosmos Reason vision-language models, or fine-tune Black Forest Labs’ FLUX.1 image models directly on their desktops. NVIDIA Launches DGX Spark
Industry Support and Early Adopters
Top technology leaders — including Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI — are launching their own DGX Spark systems, while partners such as Google, Meta, Microsoft, Hugging Face, JetBrains, Anaconda, Docker, LM Studio, Ollama, and Roboflow are testing and validating software optimized for this new platform. NVIDIA Launches DGX Spark.
Academic research groups like NYU’s Global Frontier Lab have already previewed DGX Spark. Professor Kyunghyun Cho praised the system, noting that desktop-level petascale computing makes it easier than ever to prototype and experiment with advanced AI algorithms, even for privacy-sensitive fields such as healthcare.
Availability and Ordering
Beginning Wednesday, October 15, NVIDIA DGX Spark is available to order through NVIDIA.com and from select global partners. In the United States, systems will also be sold through Micro Center and major OEM partners.
With DGX Spark, NVIDIA aims to democratize high-performance AI computing—empowering a new generation of developers, researchers, and innovators to push AI boundaries from their own desktops.
Leave a Reply