In June 2026, NVIDIA made waves at COMPUTEX 2026 by officially launching the RTX Spark superchip, a move that signaled its ambitious foray into the consumer PC processor arena. Developed in close partnership with Microsoft and crafted using TSMC’s cutting-edge 3nm manufacturing process, this innovative chip seamlessly integrates a 20-core Grace CPU with a Blackwell RTX GPU. Packed with 6,144 CUDA cores, it supports up to 128GB of unified memory, delivers an impressive 1 petaFLOP of AI computing power, and empowers local execution of massive 200B-parameter AI models.

During Clang compilation tests, the RTX Spark shone brightly, scoring 43,149—a performance that surpasses the Apple M5 by a notable 54.13%, though it narrowly missed the mark set by the M5 Pro. When it came to multi-threaded compilation, the chip’s 20-core architecture proved to be a game-changer, showcasing significant advantages and demonstrating superior energy efficiency compared to traditional x86 architectures.

Leading manufacturers such as ASUS and Dell are set to roll out laptops featuring the RTX Spark this fall, while Microsoft is simultaneously fine-tuning Windows to better support local AI task execution. The arrival of the RTX Spark marks a pivotal moment in the evolution of AI computing, shifting power from the cloud to edge devices and fueling demand growth across the communication hardware supply chain—spanning optical modules, switches, and beyond.