Microsoft and NVIDIA Tease Local AI Event Featuring RTX Spark Platform
Microsoft and NVIDIA announced an upcoming online developer event on October 7 focused on how local AI will shape the next generation of PCs. Featuring Microsoft CEO Satya Nadella and NVIDIA CEO Jensen Huang, the event will highlight NVIDIA's RTX Spark platform and new Surface hardware. This collaboration signals a major push toward executing large language and AI models natively on personal computers rather than relying heavily on cloud infrastructure. Bringing high-bandwidth unified memory and full CUDA support to Windows on Arm could significantly boost local AI capabilities for developers and power users. NVIDIA's RTX Spark superchip architecture integrates a Grace Arm CPU with a Blackwell GPU on a single substrate using NVLink-C2C interconnect technology and shared LPDDR5X memory. The upcoming Surface Laptop Ultra is expected to support up to 128GB of unified memory, enabling it to run AI models with up to 120 billion parameters directly on-device.
## BACKGROUND
AI PC initiatives aim to run machine learning models directly on end-user hardware, delivering lower latency, offline capability, and enhanced privacy. NVIDIA's NVLink-C2C is a high-speed, cache-coherent chip-to-chip interconnect designed to allow CPUs and GPUs to share memory pools efficiently at high bandwidths.