RTX Spark Live Demos: See NVIDIA's Laptop Super-Chip In Action

HotHardware attended a series of live demos showcasing NVIDIA's RTX Spark super-chip in action on Microsoft Surface Laptop Ultra systems at a special press event following the Computex 2026 unveiling. The demonstrations — spanning gaming, creative development, and on-device AI inference — provided the first extended look at what NVIDIA's Arm-based super-chip can actually do in real-world scenarios. Formerly codenamed N1 and N1X, RTX Spark combines NVIDIA's Blackwell-based GPU architecture with a 20-core Grace CPU, an NVLink C2C interconnect, and up to 128GB of high-speed unified memory. Through partnerships with Microsoft, MediaTek, and six major OEMs, NVIDIA is hoping to redefine the boundaries of mobile computing for the agentic AI era — and the live demos offered a compelling preview of that vision in practice.

The first demo highlighted the chip's gaming credentials, running Alan Wake 2 with full path tracing on an RTX Spark-powered Surface Laptop Ultra. Crucially, the demo provided an early glimpse of NVIDIA's upcoming DLSS 4.5 Ray Reconstruction update — on track to arrive in August — running on the Arm-based silicon. HotHardware noted that it's impressive to see RTX Spark already running DLSS 4.5 Ray Reconstruction at this early stage, confirming that the super-chip will support all of the latest RTX technologies right out of the gate. The second demo showcased a workflow that traditional consumer mobile GPUs cannot replicate: loading an intricately detailed Unreal Engine city project — 80GB of data — entirely into the unified memory pool in real time. In a conventional laptop environment, game developers are restricted by available VRAM, often forced to build environments in smaller, unlit blocks. RTX Spark enables developers to work inside a fully lit, high-fidelity world without the typical memory constraints.

The final demo shifted focus to on-device generative AI, leveraging the super-chip's large pool of unified memory to run a heavy 35-billion-parameter Qwen 3.635B model locally on 60–70GB of memory. The system demonstrated how RTX Spark can offload repetitive development Q&A tasks that bog down projects, running entirely on-device without any cloud dependency. Between the big pool of unified memory, next-gen DLSS 4.5 capabilities, and the robust architecture partnerships with Microsoft, MediaTek, and six OEMs, HotHardware concluded that NVIDIA's ambitious goal of reinventing the Windows PC for the age of personal AI looks attainable — and that what they saw in the live demos was certainly impressive, positioning RTX Spark as a genuine new category of mobile computing ahead of the Fall 2026 launch.