Nvidia’s RTX Spark chip moves into mini PCs from Asus, Dell, and MSI
NVIDIA official newsroom (nvidianews.nvidia.com/bios/jensen-huang)
The compact desktops pair a Blackwell GPU with a 20-core Grace CPU and start in the mid-$4,000 range
Nvidia’s RTX Spark platform now comes in a desktop version. After debuting in laptops earlier this year, the RTX Spark N1X chip is shipping inside mini PCs from Asus, Dell, and MSI.
Desktops from Acer, Gigabyte, and HP are expected to follow. Microsoft’s Surface RTX Spark Dev Box is slated to start shipping this month.
What’s inside the box
Every one of the new mini PCs runs on the same core silicon. Each pairs a Blackwell-based RTX GPU with 6,144 CUDA cores and a 20-core Nvidia Grace CPU.
Memory tops out at 128GB of unified LPDDR5X. Unified means the CPU and GPU draw from one shared pool instead of guarding separate stashes. For AI work, that matters, because large models need a lot of room to sit in memory at once.
Nvidia rates the systems at 1 petaflop of AI performance. A petaflop is a quadrillion calculations per second.
The desktops can run local models with up to 120 billion parameters. Running one that size locally means prompts and data never have to leave the building.
The physical footprint is modest. The designs measure 150x150x51 mm, which works out to roughly 1.1 liters of volume.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
That small chassis still allows a power envelope of up to 140W TDP. The systems also include 10Gb Ethernet, a meaningful upgrade over the gigabit ports found on most consumer desktops.
The price of local AI
All that hardware carries a matching price tag. Asus’s ProArt GR1X starts at $4,499 for a build with 64GB of memory and 1TB of storage.
MSI’s version comes in slightly lower, starting at $4,399 following a recent price adjustment.
For comparison, RTX Spark laptops launched on October 16, 2026, at a starting price of $2,599.
Background: Nvidia’s push into personal AI machines
Nvidia built its reputation on graphics cards for gamers, then on the data center chips powering much of the current AI boom. RTX Spark represents a different bet: a complete CPU-and-GPU platform designed for personal computers.
Native CUDA support is a core part of the pitch. CUDA is Nvidia’s programming framework, and much of the AI software world is built around it. A developer prototyping on one of these boxes is working in the same environment their code would likely run on in a data center.
What this means
Privacy is another selling point. Companies handling sensitive data may prefer to keep model inference inside their own walls. A 120-billion-parameter model running on a desk removes a third party from the equation.
For PC makers, the rollout spreads Nvidia’s reference design across a wide roster of brands. Asus, Dell, MSI, Acer, Gigabyte, HP, and Microsoft are all involved at launch or soon after.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

