Nvidia Rebrands Project DIGITS as DGX Spark: The $3,000 Personal AI Supercomputer Goes Mainstream

One year after stunning the tech world with Project DIGITS—a desktop-sized AI supercomputer priced at just $3,000—Nvidia took the concept mainstream at CES 2026. The company announced that Project DIGITS would be rebranded as DGX Spark, and that major PC manufacturers including ASUS, Dell, HP, and Lenovo would build their own versions of the device. The move signals Nvidia's intent to make personal AI supercomputing as ubiquitous as the graphics cards that built the company's empire.

The announcement came alongside the unveiling of DGX Station, a higher-performance desktop supercomputer powered by Nvidia's upcoming Blackwell Ultra platform. Together, the two products define a new category: desktop AI systems capable of running the same models that previously required cloud computing or massive data center installations.

What Is DGX Spark?

DGX Spark—the production name for what was previously called Project DIGITS—is a compact computer about the size of a Mac Mini that packs capabilities previously available only to deep-pocketed corporations and research institutions. At its heart is the GB10 Grace Blackwell Superchip, a custom silicon package that combines Nvidia's latest Blackwell GPU with an Arm-based Grace CPU.

"We're putting a petaflop of AI computing power on every desk and at every AI developer's fingertips. This is the democratization of artificial intelligence."

— Jensen Huang, Nvidia CEO

Key specifications of DGX Spark:

• Processing power: 1 petaflop (1,000 trillion operations per second) of AI compute
• Memory: 128GB of unified coherent memory, shared between CPU and GPU
• Storage: Up to 4TB of NVMe solid-state storage
• Model capacity: Can run large language models up to 200 billion parameters locally
• Power requirements: Runs on a standard electrical outlet (no specialized power needed)
• Networking: ConnectX technology allows linking two units for 405-billion parameter models
• Price: Starting at $3,000 when it ships in May 2026

The Partner Ecosystem

Perhaps more significant than the rebranding is Nvidia's decision to open DGX Spark to third-party manufacturers. ASUS, Dell, HP, and Lenovo will all produce their own versions of the device, each potentially adding their own design flourishes, cooling solutions, and support infrastructure.

This partner strategy mirrors how Nvidia built its graphics card dominance: by providing the core technology while allowing partners to compete on implementation, pricing, and service. For enterprise customers, buying from established vendors like Dell or HP provides familiar procurement processes, support contracts, and warranty terms.

The partners bring distinct advantages:

• Dell: Enterprise relationships and global support infrastructure
• HP: Strong presence in corporate IT departments
• Lenovo: Scale and aggressive pricing capability
• ASUS: Enthusiast credibility and innovative designs

DGX Station: The Premium Option

For developers and researchers who need more power than DGX Spark provides, Nvidia introduced DGX Station—a desktop workstation built around the upcoming Blackwell Ultra platform. While pricing wasn't disclosed, DGX Station is expected to cost significantly more than the entry-level Spark but offer proportionally greater capability.

DGX Station targets users who need to work with the largest AI models locally, including those developing new foundation models or fine-tuning existing ones with massive datasets. The system essentially brings data center-class capabilities into an office environment.

Who Will Buy These Devices?

Nvidia has identified several target audiences for the DGX Spark and DGX Station lineup:

AI Developers and Researchers: The primary use case is prototyping and developing AI applications. Instead of paying hourly cloud computing costs, developers can work on their own hardware with no usage limits. For active developers, the $3,000 investment pays for itself quickly compared to cloud alternatives.

Data Scientists: Analysts who work with large datasets and machine learning models can run inference locally, maintaining data privacy and avoiding cloud transfer costs.

Students and Academics: Universities and individual students can access cutting-edge AI capabilities without institutional data center resources.

Enterprise IT: Companies can deploy DGX Spark units for edge computing, local inference, and AI development that requires data sovereignty.

The Competitive Landscape

DGX Spark enters a market that barely existed before Nvidia created it. While Apple's M-series Macs offer impressive on-device AI capabilities, they lack the raw power needed for serious model development. AMD has hinted at competing products but has yet to deliver anything comparable.

Cloud providers like AWS, Google Cloud, and Microsoft Azure represent indirect competition—users can rent Nvidia GPUs by the hour rather than buying dedicated hardware. However, for developers who work with AI daily, ownership economics often favor local hardware.

Investment Implications

The DGX Spark announcement reinforces several investment themes:

• Nvidia's ecosystem expansion: By moving beyond data centers to desktops, Nvidia addresses a new market segment
• Partner stock catalyst: Dell, HP, and Lenovo shares could benefit from the association with Nvidia's AI platform
• Developer tool spending: Companies investing in AI will increasingly provision developer workstations with high-end AI capabilities

For Nvidia specifically, DGX Spark represents a relatively small revenue opportunity compared to data center sales—each unit generates only $3,000 compared to hundreds of thousands for enterprise GPU systems. However, the strategic value is enormous: every developer who learns on Nvidia hardware becomes an advocate for Nvidia solutions in their eventual enterprise deployments.

Looking Ahead

The rebranding from Project DIGITS to DGX Spark and the expansion to partner manufacturers mark the transition from concept to commercial product. When units begin shipping in May, they'll represent the first mass-market AI supercomputers—devices that were science fiction just a few years ago.

For the AI industry, the implications are profound. Democratizing access to powerful AI development tools could accelerate innovation, as more developers experiment with models and applications that were previously possible only at large organizations. The next breakthrough AI application might well be prototyped on a DGX Spark sitting on someone's desk.

Jensen Huang summarized the vision: "We built the GPU that powered the gaming revolution, and then the data center revolution. DGX Spark is how we'll power the personal AI revolution."