NVIDIA Unveils RTX Spark, Its First Consumer PC Superchip

NVIDIA is moving deeper into the consumer PC market with RTX Spark, a new system-on-chip designed to bring its AI, graphics and creator technologies into a single Windows-based platform.

Unlike a traditional GPU launch, RTX Spark is being positioned as a full PC foundation. The chip combines a 20-core NVIDIA Grace CPU with a Blackwell RTX GPU, fifth-generation Tensor Cores, up to 128GB of unified memory and up to 1 petaflop of AI performance. In practical terms, NVIDIA wants this to power thin laptops and compact desktops capable of local AI work, advanced creative tasks and high-end gaming without relying entirely on cloud infrastructure.

The company is also working closely with Microsoft to make RTX Spark part of a broader push toward “personal AI agents” on Windows. These systems are being designed to run AI tools locally, handle complex workflows, protect user data more effectively and support agent-based features directly inside the Windows environment.

For creators and developers, NVIDIA is promising serious performance: editing 12K video, rendering very large 3D scenes, generating AI video, running large language models locally and accelerating major apps such as Photoshop, Premiere, Blender, DaVinci Resolve and ComfyUI. Gaming is also part of the pitch, with support for RTX technologies including DLSS, ray tracing, Reflex and G-SYNC.

RTX Spark-powered laptops and compact desktops are expected this fall from ASUS, Dell, HP, Lenovo, Microsoft Surface and MSI, with Acer and GIGABYTE models planned later. The bigger question is not whether the chip is ambitious - it clearly is - but how well software support, pricing and real-world battery performance match NVIDIA’s claims once devices reach consumers.