Microsoft is making waves with its announcement to bring distilled DeepSeek R1 AI models to Windows 11 Copilot+ PCs, starting with Snapdragon X powered devices. This development promises to revolutionize the way developers create AI-powered applications by enabling them to run directly on-device. The initiative follows the recent introduction of DeepSeek R1 on Azure AI Foundry, a platform designed for leveraging advanced AI tools and machine learning models.
Initially, these AI models will be tailored for Snapdragon X laptops, tablets, and PCs. However, Microsoft plans to expand this technology to systems powered by Lunar Lake and AMD Ryzen AI 300 processors. Through its Developer Blog, the company noted that Neural Processing Unit (NPU) optimized versions of DeepSeek-R1 will soon be available on Copilot+ PCs, including those from top manufacturers like Asus, Lenovo, Dell, HP, and Acer. These systems boast processors with impressive NPU capabilities, supporting 40 to 45 TOPS (Tera Operations Per Second).
Developers can expect to receive the first iteration, DeepSeek-R1-Distill-Qwen-1.5B, via the AI Toolkit shortly, with more advanced versions like the 7B and 14B in the pipeline. This advancement will empower developers to effectively build and deploy AI-driven applications that fully utilize the hardware’s processing potential.
While Microsoft hasn’t specified an exact rollout schedule, Intel Core Ultra 200V devices are anticipated to gain access to these AI enhancements sooner than AMD Ryzen AI models. In the interim, AMD is proactively equipping developers with guides related to DeepSeek R1 Distill for its latest Ryzen AI 300 series APUs, as well as the Ryzen 8040/7040 chips, signalling their commitment to integrating AI capabilities despite their relatively lower NPU performance of 16/10 TOPS, respectively.
