On January 29, Microsoft announced that it has integrated the R1 artificial intelligence model from Chinese startup DeepSeek into its Azure cloud computing platform and GitHub tool for developers.
The AI model will be added to the model catalog on these platforms, expanding Microsoft's collection to over 1,800 models. DeepSeek claims that its model is more cost-effective and uses less data than existing services, leading to its rapid adoption over U.S. competitor ChatGPT on Apple's App Store.
This strategic move aims to lessen Microsoft's reliance on ChatGPT's maker, OpenAI, by incorporating a variety of AI models, including third-party options, into its Microsoft 365 Copilot flagship AI product. Microsoft plans to enable customers to run the R1 model locally on their Copilot+ PCs to address privacy and data-sharing concerns related to the model's usage.
DeepSeek's storage of user data on servers in China may pose challenges to its adoption in the U.S. market. Meanwhile, Microsoft and OpenAI are investigating whether data output from OpenAI was inappropriately accessed by a group associated with DeepSeek.
The emergence of DeepSeek in the AI landscape has prompted competitors to react, with OpenAI announcing a specialized version of ChatGPT for U.S. government agencies. In a related move, China's Alibaba released a new Qwen 2.5 AI model, coinciding with the Lunar New Year.