- Nvidia announced the RTX Spark, an Arm-based superchip combining a 20-core Grace CPU and Blackwell GPU with up to 128 GB of unified memory.
- The chip delivers 1 petaflop of local AI performance in a 45W to 80W power envelope, targeting laptops and desktops shipping this fall.
- Microsoft, Dell, HP, ASUS, Lenovo, and MSI will build the first wave of RTX Spark devices, with Microsoft's Surface Laptop Ultra among them.
Nvidia bets its data center dominance can translate to personal computing
Jensen Huang used his two-hour COMPUTEX 2026 keynote in Taipei on June 1 to make the company's boldest move outside the data center in years. The RTX Spark, which Nvidia also calls the N1X, pairs a 20-core Grace CPU with a Blackwell GPU packing 6,144 CUDA cores, all on TSMC's 3-nanometer process. The two processors share a single pool of up to 128 GB of LPDDR5X memory, eliminating the bottleneck that separates CPU and GPU memory in conventional laptops.
The integrated GPU is roughly equivalent to a standalone RTX 5070. Nvidia claims the chip can handle 1440p gaming at 100 frames per second with DLSS 4.5 upscaling, while simultaneously delivering 1 petaflop of AI compute for on-device model inference. The power envelope sits between 45 and 80 watts, competitive with Apple's M-series and Qualcomm's Snapdragon X Elite.
Six OEMs confirmed launch hardware for fall 2026, with Acer and GIGABYTE expected to follow. Microsoft co-developed the platform and will ship a Surface Laptop Ultra variant. Nvidia is also working with Adobe to rebuild Photoshop's core as a fully GPU-accelerated application optimized for RTX Spark.
"Every token is now a revenue unit."
Jensen Huang, Nvidia CEO, COMPUTEX 2026 keynoteWhy the PC play matters for Nvidia's AI strategy
The keynote's central argument was that agentic AI needs to run locally, not just in the cloud. Huang announced the DGX Station for Windows alongside RTX Spark, a workstation capable of running models with up to 1 trillion parameters on-device. Combined with Microsoft's Windows Local AI runtime shipping June 9, the pitch is a full stack for developers who want AI agents running on consumer and professional hardware without round-tripping to a data center.
Huang called RTX Spark "the most stunning product in Nvidia's 33-year history" and said the company wants to "reinvent the single most important tool of humanity." For Nvidia, the personal computer is no longer a sideshow to the data center business. If RTX Spark delivers on its specifications at competitive pricing, Nvidia will have planted a flag in the one major compute market where it has never been a primary player.
The broader COMPUTEX keynote also confirmed that the Vera Rubin data center platform is in full production, with a supply chain twice the size of Grace Blackwell, and introduced Cosmos 3, an open world foundation model for physical AI. But RTX Spark was the headline: Nvidia turning its attention from training clusters to the laptop on your desk.
Santage is committed to independent, transparent journalism. This article is produced in accordance with Santage's Editorial Standards and aims to provide accurate and timely information. Readers are encouraged to verify information independently.