Skip to content
QbitAI · WeChat

Jensen Huang Brings NVIDIA CPUs Into the PC Market

老黄带着英伟达CPU杀入PC赛道

NVIDIA RTX Spark will ship in Windows PCs this fall with 1 petaflop of AI compute and 128GB unified memory. The platform combines a Blackwell RTX GPU, a 20-core Arm-based Grace CPU, and NVLink-C2C, and NVIDIA says it can run 1-million-token-context, 120B-parameter language models locally.

Why it matters: NVIDIA packs a 20-core Grace CPU and a Blackwell GPU into a Windows machine over NVLink-C2C with 128GB of shared unified memory, a hardware path to running large models locally that differs from discrete GPUs.

Read the original ↗Export Markdown