
Microsoft is testing a new experimental feature in the Windows App SDK that lets developers run local language models on non-Copilot+ PCs using an Nvidia GPU. Supported hardware includes Nvidia GeForce RTX 30 series and newer with 6GB or more vRAM. This expands local AI capabilities like text rewriting and summarizing beyond NPU-equipped Copilot+ PCs.
Tap to vote and see what everyone thinks.