NVIDIA on June 1 unveiled the RTX Spark Superchip, a Windows AI-PC processor that combines a 20-core ARM-based Grace CPU with a Blackwell-generation RTX GPU and up to 1 petaflop of AI performance. The chip arrives this fall in laptops and desktops from Dell Technologies and Lenovo, and is already powering Microsoft's new Surface Laptop Ultra with 128 GB of unified memory.
What's in the Superchip
RTX Spark is a single-package design integrating Grace CPU cores licensed from Arm with NVIDIA's latest Blackwell graphics architecture and on-package memory. The result is a x86-class Windows experience running on ARM, with the GPU delivering peak AI performance comparable to a desktop workstation. NVIDIA is positioning the platform for agentic-AI workloads — local copilots, on-device coding agents, image and video generation — that today require cloud round trips.
OEM rollout: Dell, Lenovo, Microsoft
Dell Technologies will integrate RTX Spark across its XPS, Precision and ProArt-class workstation lines. Lenovo confirmed designs across ThinkPad, ThinkStation and Legion family devices. Microsoft's Surface Laptop Ultra is the launch flagship, shipping with 128 GB RAM and a TSMC 3N-fabricated Spark Superchip. Manufacturing of the chip itself is happening at TSMC under the broader NVIDIA-TSMC AI-in-fabs collaboration announced the same week.
Wall Street response: AMD, Intel, Qualcomm under pressure
Shares of Advanced Micro Devices, Intel and Qualcomm fell on Monday as Wall Street digested the threat. Arm Holdings traded higher on volume expectations from the Grace CPU win. Intel had only just announced its Crescent Island AI inference platform for the data center; the consumer Windows-on-Arm vector is a more direct hit on AMD's Ryzen AI roadmap and Qualcomm's Snapdragon X share gains over the last two cycles.
Strategic stakes
RTX Spark closes one of the last gaps in NVIDIA's stack: a consumer compute platform under its direct control. CEO Jensen Huang's pitch — that NVIDIA wants to own a layer at every tier of the AI stack, from foundry partnerships through cloud accelerators to humanoid robots like the Isaac GR00T reference humanoid — gains a Windows endpoint with this launch.
Availability
Devices are scheduled to ship starting fall 2026. NVIDIA has not disclosed pricing, but partners signaled premium tiers above current Snapdragon X Elite and AMD Strix Halo configurations. Enterprise buyers should expect Spark variants tied to managed AI-PC suites from Microsoft and Dell.
Reporting based on coverage from CNBC, Bloomberg, Spokesman and Advisor Perspectives.