Nvidia launches RTX Spark for AI PCs

Nvidia introduced RTX Spark, a combined CPU and GPU superchip, at Computex 2026 in Taipei, presenting it as a platform for local personal AI agents and high-end creative workloads. Per Nvidia and Microsoft, RTX Spark delivers 1 petaflop of AI performance and up to 128GB of unified memory, pairing a 20-core Arm-based Nvidia Grace CPU with a Blackwell RTX GPU carrying up to 6,144 cores connected over NVLink-C2C. Nvidia and Microsoft said multiple OEMs, including ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI, will ship RTX Spark laptops and compact desktops in fall 2026, with Acer and Gigabyte to follow.

Per Nvidia product materials and the Microsoft Windows blog, RTX Spark integrates GPU, CPU, and memory on a power-efficient design intended for thin-and-light laptops (as slim as 14mm) and small desktops. Nvidia lists workloads such as rendering 90GB+ 3D scenes, editing 12K 4:2:2 video, generating 4K AI video, and running 120B-parameter language models with up to 1 million token contexts on-device. Nvidia and Microsoft also detailed Windows security primitives and Nvidia OpenShell for on-device agents, plus optimizations for local-agent runtimes such as llama.cpp.

Reuters places the launch against several years of PC-industry attempts to market "AI PCs" that have not generated broad consumer demand, noting earlier efforts centered on features like transcription and image editing and struggled to move meaningful sales. Reuters cites analysts who say premium pricing and a memory-supply crunch could limit RTX Spark to niche adoption, and quotes Tirias Research analyst Kevin Hein describing it as creating "a new category between the workstation and the AI server" rather than replacing traditional PCs.

On-device capability at the scale Nvidia describes, a unified memory pool approaching 128GB and high inference throughput, materially expands what is feasible without cloud roundtrips, especially for privacy-sensitive or latency-critical agent workflows. As an industry pattern, though, hardware alone rarely guarantees ecosystem adoption: software porting, toolchain support, and application-level integration determine whether developers and creative apps actually exploit that on-device scale. Adobe's plan to rearchitect Photoshop and Premiere and other early app support help, but broader developer tooling and model optimization remain the gating factors.

Watch OEM pricing and configuration tiers as systems reach market, since Reuters flags price as a key adoption limiter; memory-supply and component availability that could affect lead times and cost; and independent software ports and runtime benchmarks for real workloads, including llama.cpp and model-load tests. Enterprise uptake and adoption of the new Windows security primitives will signal whether local agents are marketed mainly to developers and creators or positioned for wider consumer use.