Nvidia has unveiled a new artificial intelligence chip, RTX Spark, that can run large language models locally on Windows-based laptops.

IT outlet ITmedia reported on Sunday that Nvidia Chief Executive Jensen Huang (젠슨 황) announced RTX Spark in a keynote speech at Nvidia GTC Taipei 2026 in Taiwan.

RTX Spark is an Arm-based system-on-chip co-developed with MediaTek. Nvidia introduced the chip as targeting demand for local AI execution on laptops. It highlighted that AI models can run directly on Windows laptops without relying on the cloud.

The chip integrates a 20-core Grace CPU and a Blackwell-generation GPU. It supports up to 128GB of integrated LPDDR5X memory, with memory bandwidth reaching 600GB per second. AI performance was presented as 1 petaflop, or 1,000 trillion operations per second.

Nvidia said RTX Spark can run LLMs with up to 120 billion parameters locally on laptops. It said supported context length is up to 1 million tokens, stressing that large AI models can be handled directly on personal devices rather than cloud servers.

Its target users are creators, developers and gamers. Nvidia cited examples such as rendering 3D scenes exceeding 90GB, editing 12K 4:2:2 video, generating 4K AI video and running AAA games at 1440p resolution at more than 100 frames per second. It aims to handle generative AI and high-performance graphics workloads together on laptop-class devices.

Software support will also be provided. Adobe said it will optimise Photoshop and Premiere for RTX Spark to double AI and graphics performance. This is seen as a move to boost usefulness in actual production tools rather than only highlighting chip performance figures.

Plans to include the product were also unveiled. Microsoft said it will apply RTX Spark to its Surface lineup. Asus, Dell, HP, Lenovo and MSI also plan to launch slim laptops and small desktop PCs equipped with RTX Spark in fall 2026.

As a result, RTX Spark is increasingly likely to serve as a signal that Nvidia is stepping up its push into the Windows ecosystem in the AI PC market. With the Arm-based SoC design, large integrated memory and local LLM execution capability brought to the fore, how performance is delivered and the scope of optimisation for major applications are expected to be key points after products are released in the second half of the year.

This is the NVIDIA RTX Spark Superchip. A new beginning for personal computers. Designed for creators, AI developers, and gamers, RTX Spark brings over 30 years of NVIDIA innovation to slim Windows laptops and small, ultra-efficient desktop PCs. pic.twitter.com/RmcamHTS4z