Nvidia unveils RTX Spark, aims to bring AI agents to Windows PCs
Nvidia packed a 1-petaflop AI superchip into Windows PCs, betting local agents and on-device privacy can trigger the next upgrade wave.
Nvidia used its Computex Taipei keynote to pitch a new reason to buy a PC: not just faster apps, but machines that can run AI agents locally, securely and without sending every task to the cloud. The company unveiled RTX Spark, a 1-petaflop superchip for Windows laptops and desktops built around a Blackwell RTX GPU with 6,144 CUDA cores, a 20-core Grace CPU and up to 128GB of unified memory.
Nvidia and Microsoft framed the system as the foundation for personal AI computers and Windows-native agents. The company said the point is to put AI capabilities directly on a user’s primary device, where privacy and security can be managed more tightly than in a cloud-only setup. Nvidia has argued that broad adoption of agentic AI has been held back because people have not been able to run these systems securely and privately on the PCs they already use every day.
The hardware pitch is aimed at creators, developers and business users who need local performance, not just chat windows. Nvidia said RTX Spark can render 90GB-plus 3D scenes, edit 12K 4:2:2 video, generate 4K AI video and run 120B-parameter language models with up to 1 million tokens of context on-device. At the same time, Nvidia said the machines can still handle AAA games at 1440p and more than 100 frames per second. Adobe is rearchitecting Photoshop and Premiere for RTX Spark, which Nvidia said should deliver 2x faster AI and graphics performance.
The rollout also shows how aggressively Nvidia is moving beyond data-center GPUs into the PC market long dominated by Intel and AMD, and more recently challenged by Qualcomm and Apple. Slim Windows laptops and compact desktops using RTX Spark are set to arrive this fall from ASUS, Dell, HP, Lenovo, Microsoft Surface and MSI, with Acer and GIGABYTE to follow. MediaTek collaborated on the custom CPU design, signaling that Nvidia is leaning on a broader partner network to break into personal computing.
Jensen Huang has said Nvidia’s Vera CPU opens access to a new $200 billion market, and he said in Taipei that the forecast includes China. The scale of that opportunity explains the company’s urgency: if local AI agents become a mainstream need, RTX Spark could help drive a new PC upgrade cycle. If they do not, the launch will look like another costly bet on hardware ahead of demand.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
