Nvidia unveiled the RTX Spark superchip at Computex in Taipei on June 1, 2026, designed to run AI agents natively on Windows laptops without cloud dependency. CEO Jensen Huang described the technology as turning PCs into "collaborative teammates."
Built on the Blackwell architecture, RTX Spark delivers up to 1 petaflop of AI performance. It features a 20-core Grace CPU with FP4 support, up to 128GB of unified memory, and can run AI models with up to 200 billion parameters and 1 million token context windows entirely on-device. The chip supports Nvidia’s full CUDA/RTX software stack and includes security features co-developed with Microsoft.
ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI will ship RTX Spark devices starting fall 2026. Acer and GIGABYTE are expected to follow. Adobe is optimizing Photoshop and Premiere for the platform.
This announcement follows Nvidia’s DGX Spark supercomputer, positioning RTX Spark as the consumer-focused counterpart. Local AI processing eliminates cloud costs and latency, benefiting applications such as AI-powered trading analysis, dApp development, and crypto workflows.