Image Courtesy: Nvidia
NVIDIA has unveiled RTX Spark, a new AI-focused system-on-a-chip designed to bring high-performance artificial intelligence and graphics capabilities to Windows laptops and compact desktops. Announced at Computex, the chip is positioned as a direct competitor to emerging AI-centric processors from AMD and Qualcomm while introducing NVIDIA’s Blackwell architecture to a new category of consumer devices.
The company says RTX Spark delivers up to one petaflop of AI computing performance and combines 6,144 Blackwell RTX cores with 20 Arm-based CPU cores developed in partnership with MediaTek. The chip is expected to power devices including the Surface Laptop Ultra, Dell XPS 16, and systems from major PC manufacturers worldwide.
Unlike traditional laptop designs that pair a CPU with a separate graphics processor, RTX Spark integrates AI, graphics, and computing workloads into a single package. NVIDIA claims the chip can deliver performance comparable to an RTX 5070 laptop GPU while consuming significantly less power, with configurations ranging from single-digit wattage to 80 watts depending on the device.
A major focus of the platform is artificial intelligence. While the chip includes a neural processing unit capable of meeting Microsoft’s Copilot+ PC requirements, NVIDIA is emphasizing the AI capabilities of its Blackwell GPU architecture and tensor cores. The company says the chip can access a large pool of unified memory ranging from 16GB to 128GB, allowing AI models and graphics workloads to share resources more efficiently.
NVIDIA CEO Jensen Huang described the technology as part of a broader transformation of personal computing, where AI agents could eventually handle many tasks that currently require direct user interaction. The company has reportedly spent several years working alongside Microsoft to optimize Windows for the new architecture.
Microsoft has also confirmed operating system-level enhancements for the platform. The company says Windows 11 has been optimized to intelligently schedule workloads across the chip, improving both performance and power efficiency for everyday tasks and AI-driven applications.
The launch highlights NVIDIA’s ambition to expand beyond discrete graphics cards and establish a larger presence in the rapidly growing AI PC market. While the company’s DGX Spark desktop targets enterprises and AI developers, RTX Spark is aimed at mainstream creators, developers, and consumers who want local AI processing without relying entirely on cloud services.
The move also marks NVIDIA’s return to consumer-facing system-on-chip development, a market it largely exited after its Tegra processors. This time, however, the company enters a much more competitive landscape where AI capabilities are becoming one of the primary battlegrounds for the future of personal computing.