Microsoft launches the AI model from DeepSeek on Azure

Microsoft announced on Wednesday that it has made the R1 artificial intelligence model from Chinese firm DeepSeek available on its Azure cloud computing platform and developer tool, GitHub.

The AI model will join the more than 1,800 models that Microsoft now offers and be accessible in the model catalog on the platforms.

Last week, DeepSeek released a free AI assistant that, according to the company, utilizes less data and costs a small fraction of what other companies charge. Investors in IT stocks were alarmed when the assistant surpassed ChatGPT, a rival in the United States, in downloads from Apple’s App Store on Monday.

Microsoft made the move in an effort to lessen its reliance on OpenAI, the company that created ChatGPT. According to a Reuters story from last month, the business has been striving to integrate both internal and external AI models to power its main AI product, Microsoft 365 Copilot.

Additionally, Microsoft announced that users will soon be able to run the R1 model locally on their Copilot+ PCs, which may allay worries about data sharing and privacy.

DeepSeek has stated that it keeps user data on Chinese servers, which may be a barrier to its adoption in the United States.

Bloomberg News reported on Tuesday that Microsoft and OpenAI are investigating if data output from OpenAI’s technology was illegally taken by a party associated with DeepSeek.

With OpenAI CEO Sam Altman stating that the business will “pull up some releases” in response to DeepSeek’s explosive rise in the AI market, the company on Tuesday unveiled a customized version of ChatGPT for U.S. government agencies.

The release of a new version of Alibaba’s Qwen 2.5 AI model on Wednesday was also unexpected given that it was the first day of the Lunar New Year.