Microsoft is closely associated with OpenAI’s ChatGPT AI model, yet the tech giant is not hesitant to explore other opportunities. Recently, Microsoft announced its plans to integrate the DeepSeek-R1 AI model into Copilot+ PCs, commencing with Snapdragon X devices and subsequently extending to Intel Lunar Lake and AMD Ryzen AI 9 PCs. The DeepSeek-R1-Distill-Qwen-1.5B model is expected to launch “soon” on the Microsoft AI Toolkit for developers, with more powerful 7B and 14B variants scheduled for future release.
While the 1.5B (base) model may not match the performance of the higher-tier 32B and 70B models, Microsoft emphasizes that these models are “NPU-optimized” for Copilot+ PCs. The minimum specifications required for such devices include 256GB of storage, 16GB of RAM, and an NPU capable of at least 40 TOPS (trillions of operations per second).
Microsoft noted, “These optimized models enable developers to build and deploy AI-powered applications that run efficiently on-device, leveraging the advanced NPUs in Copilot+ PCs.” The company also implemented systems designed to utilize low-bit processing, ensuring that the R1 models can operate locally on NPU hardware.
Simultaneously, Microsoft is set to introduce DeepSeek’s R1 model to its Azure AI Foundry platform, as reported by The Verge. This addition will place it alongside other AI models on the platform, including GPT-4, Mistral AI, Meta-Llama 3, and more. This development comes as a surprise, especially considering that Microsoft is reportedly investigating whether DeepSeek improperly utilized OpenAI’s technology.
Concerns regarding privacy issues surrounding the China-based DeepSeek have also been raised, which Microsoft addressed in a Marketplace Community post. According to Microsoft senior product marketing manager, Justin Royal, “DeepSeek R1 has undergone rigorous red teaming and safety evaluations, including automated assessments of model behavior and extensive security reviews to mitigate potential risks.”
DeepSeek has made a significant impact on the AI landscape with its R1 model, which demands considerably less computing power than its competitors. This innovation has led to market fluctuations, including a selloff of shares in chip manufacturer NVIDIA and other companies in the AI sector. Recently, OpenAI has faced multiple lawsuits from newspapers and publishers globally for copyright infringement, alleging that DeepSeek and other Chinese AI startups have been “distilling” its models.
