Nvidia launches $3K desktop AI supercomputer with petaflop performance

Nvidia has launched its DGX Spark “personal AI supercomputer” for sale this week, priced at $3,000 and designed to fit on a desktop while delivering data center-level AI performance. The compact machine democratizes access to sophisticated AI computing power that was previously only available through expensive, energy-intensive data centers, potentially transforming how researchers, students, and data scientists engage with AI development.

What you should know: The Spark delivers enterprise-grade AI capabilities in a desktop form factor that can run on standard power outlets.

• The system features Nvidia’s GB10 Grace Blackwell Superchip, 128GB of unified memory, and up to 4TB of NVMe SSD storage.
• It can deliver a petaflop of AI performance — capable of one million billion calculations per second — and handle AI models with up to 200 billion parameters.
• Nvidia calls it “the world’s smallest AI supercomputer,” emphasizing its compact desktop footprint.

How to get it: Spark is available for purchase starting Wednesday through multiple channels.

• Orders can be placed directly at nvidia.com as well as through select partners and retail stores in the US.
• Third-party manufacturers including Asus, Dell, and HP are developing their own versions of the system.
• A larger variant called Station exists, though Nvidia hasn’t announced general availability plans for that model.

Why this matters: The launch represents a significant shift in making high-performance AI computing accessible to individual researchers and smaller organizations.

• Previously, this level of AI performance required access to expensive data centers with high energy demands.
• The democratization of AI supercomputing could accelerate research and development across academic institutions and smaller companies.
• As Nvidia CEO Jensen Huang explained when first announcing the system: “placing an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to engage and shape the age of AI.”

The bigger picture: Spark positions Nvidia to capture the growing market for personal AI computing as demand for local AI processing increases amid privacy concerns and latency requirements for real-time applications.