Nvidia launches desktop AI supercomputers with Blackwell chips starting at $3,000

Nvidia unveils two new “personal AI supercomputers” powered by its Blackwell platform, offering unprecedented local AI processing power for developers and researchers. These systems enable users to work with large AI models without requiring constant datacenter connections, representing a significant advancement in bringing enterprise-level AI capabilities to desktop form factors. This launch highlights the growing competition in local AI hardware as companies race to deliver powerful, memory-rich systems for running large language models without cloud dependencies.

The big picture: Nvidia has formally introduced the DGX Spark and DGX Station personal AI supercomputers at its GTC conference, with the Spark available for preorder immediately.

• The DGX Spark, previously announced as “Digits” at CES, is priced at $3,000 and features a Mac Mini-sized form factor.
• The larger DGX Station (price not yet disclosed) targets AI developers, researchers, and data scientists who need to prototype and fine-tune large models on desktop systems.

Key specs: Both systems leverage Nvidia’s new Grace Blackwell platform, offering unprecedented AI compute power in desktop form factors.

• The Spark uses the GB10 Blackwell Superchip with fifth-generation Tensor Cores and FP4 support, delivering up to 1,000 trillion operations per second (TOPS) of AI compute.
• The DGX Station houses the more powerful GB300 Blackwell Ultra desktop superchip, providing 20 petaflops of AI performance and 784GB of unified system memory.

Hardware details: Nvidia’s systems feature substantial memory and storage configurations essential for handling large AI models locally.

• The Spark includes 128GB of unified memory and up to 4TB of NVMe SSD storage.
• Both machines support working with advanced AI models, including Nvidia’s Cosmos Reason world foundation model and GR00T N1 robot foundation model.

Market expansion: Nvidia is partnering with major OEMs to expand the availability of these systems.

• Asus, Dell, HP, Boxx, Lambda, and Supermicro will build DGX Station versions available later this year.
• Asus, Dell, HP, and Lenovo will manufacture versions of the DGX Spark, with deliveries expected this summer.

Industry context: The launch occurs amid growing competition in the personal AI hardware space.

• AMD offers the competing Ryzen AI Max+ “Strix Halo” GPU with similar capabilities.
• HP and Framework are already incorporating AMD’s solution into systems, with the latter offering a $2,000 desktop that can access up to 96GB of VRAM.