Nvidia Launches Integrated CPU-GPU Chip Aimed at Running AI Models on Personal Computers

Nvidia unveiled the RTX Spark superchip on Monday, combining central and graphics processing units in a single package designed to run advanced artificial intelligence functions directly on laptops and desktop computers. Jensen Huang, founder and chief executive of the Santa Clara, California-based company, made the announcement during the annual Nvidia GTC event in Taipei.

He said Microsoft and Nvidia “are going to reinvent the PC” and described the new machines as “AI personal computers” for creating and gaming.

The superchip will allow the computers to run AI agents locally, supporting highly capable models and complex workloads without relying on remote servers, Microsoft stated in a separate release. Huang said the devices would let users talk to an autonomous agent that understands them, ask it to read files, or request research assistance.

New personal computer models from Microsoft, Dell and other brands are expected to debut in the fall.

Nvidia also said its new Vera CPUs for data centers have entered full production and will serve as a major growth driver tied to the expansion of AI agents. Early customers are expected to include Anthropic, OpenAI and SpaceXAI. Nvidia is currently the world’s most valuable company, ahead of Apple, Google’s parent Alphabet and Microsoft.