RTX Spark laptops in Nvidia launch artwork released on October 16, 2026 (illustrative). Image: Nvidia
Nvidia’s first Windows laptops built around its own chip went on sale around the world on Friday, and the company is pitching them as PCs that run AI agents and large AI models without the cloud. Nvidia said RTX Spark laptops from ASUS, Dell, HP, Lenovo, Microsoft and MSI are available today, starting at $2,599.
The machines were first shown off at Microsoft’s Windows and Surface event on October 7, where Microsoft also set out plans for Copilot to act on the files on your PC. Friday is the day they reach shops, with Nvidia hosting launch events at stores including a Best Buy in San Francisco.
What RTX Spark is
RTX Spark is a single “superchip” that puts an Nvidia Blackwell RTX graphics chip and an Nvidia Grace processor, an Arm design built with MediaTek, side by side with up to 128GB of memory shared between them. Nvidia says it delivers up to 1 petaflop of AI performance (measured in FP4, a low-precision format used to run AI models quickly).
The shared memory is the point. A large AI model has to fit in memory to run, and on most laptops the graphics chip gets only a small slice of its own. Nvidia says RTX Spark can run “frontier AI models locally” while a creative app runs alongside, and that the same machine can edit 8K video, render 3D scenes of up to 90GB and play games at 1440p at more than 100 frames per second. Those are Nvidia’s figures; independent reviews are only starting to appear.
Jensen Huang, Nvidia’s founder and chief executive, framed it as a PC built for the AI helpers people will use next:
Together with Microsoft and our PC industry partners, we’re bringing creators, developers and gamers a new generation of Windows PC, built for the work they do today and the personal AI agents they’ll work with next.
Jensen Huang, founder and CEO, Nvidia
AI agents out of the box
Nvidia says guided installers let buyers set up AI agents without touching a command line, naming GitHub Copilot, Hermes, OpenClaw, Perplexity, Pi and WorkBuddy. It is also releasing Nemotron 3.5 Bolt, a 74-billion-parameter model of its own tuned for calling tools and handling multi-step tasks, for supported configurations. Open models built for agents, such as Meta’s Muse Glimmer, are the kind of software this hardware is aimed at.
To keep agents in check, Nvidia points to its OpenShell secure runtime and Microsoft Execution Containers, which control what an agent is allowed to touch. A feature called Nvidia PAIR lets an RTX Spark laptop serve AI to other Windows, Linux and Mac computers on a home network.
Nous Research, which makes the Hermes agent, said running locally gives users more say over their data:
We brought Hermes to RTX Spark so people can put open AI models and agents to work locally on a Windows laptop, with greater control over their data and the freedom to shape AI around their own workflows.
Dillon Rolnick, CEO, Nous Research
For developers, Nvidia and Microsoft say they have readied the first PyTorch build with CUDA for Windows on Arm, a native Python and Conda installer, and support for 97% of the top 15,000 Python packages. Adobe is shipping beta versions of Photoshop and Premiere with native RTX Spark support, with Substance 3D to follow.
Which laptops, and what they cost
Nvidia says the laptops are on sale globally through the makers’ own websites and major retailers, with pricing and configurations varying by region. The launch line-up is:
- ASUS: ProArt P14 and ProArt P16
- Dell: XPS 16
- HP: OmniBook Ultra and OmniBook X 14
- Lenovo: Yoga Pro 9n
- Microsoft: Surface Laptop Ultra, from $2,599
- MSI: Prestige N16 Flip AI+ and Prestige N16
Nvidia hasn’t published prices for each model or country; check the maker’s site in your region for local pricing in pounds, euros, Canadian or Australian dollars. Microsoft has said its Surface RTX Spark Dev Box, a $5,999 mini PC for developers, ships in November in the US only, and Nvidia says more systems, including compact desktops, will follow. New Perplexity users get a limited-time offer with the laptops.
Why it matters
RTX Spark is Nvidia’s biggest push yet to move AI from the data centre onto the desk, and it gives Windows buyers a rival to Apple’s Macs, which also share one big pool of memory and have become a popular way to run large models at home. Whether buyers pay $2,599 and up for that depends on how well local agents work in daily use, which reviews will now test.
Sources: Nvidia newsroom, Nvidia blog.
Latest Chips news
