Nvidia (NVDA) Unveils Vera CPU, RTX Spark Superchip, Challenging Wintel Alliance

Nvidia (NVDA) CEO Jensen Huang on June 1, 2026 launched the company’s first custom Arm-based CPU for AI agents, a PC superchip integrating CPU, GPU and NPU, and a robotics platform at GTC Taipei, moving aggressively into traditional CPU territory held by Intel and AMD and threatening the long-standing Wintel computing model. The Vera CPU is purpose-built for AI agents, completing tasks 1.8 times faster than traditional x86 chips, and is already in production for early adopters OpenAI, Anthropic and SpaceX. The PC flagship, RTX Spark (N1X), co-developed with MediaTek and Arm on TSMC’s 3nm process, packs a Grace CPU, Blackwell GPU with 6,144 CUDA cores, and an AI NPU into a single package with 128GB unified memory, enabling laptops to run hundred-billion-parameter AI models locally. Microsoft supports Windows on Arm, with Dell, Lenovo, ASUS, HP and MSI expected to ship devices starting this fall. Intel countered the same day with its Xeon 6+ server CPU on the 18A process, heating up direct competition. Analysts said Nvidia’s strategy—spanning data-center CPUs, AI PCs and robotics reference designs—seeks full-stack computing dominance, realigning the global computing landscape.