Microsoft MSFT.O has announced the launch of the R1 artificial intelligence model from Chinese startup DeepSeek on its Azure cloud platform and GitHub for developers, as stated by the U.S. company on Wednesday.

This AI model will be part of the model catalog on both platforms, adding to Microsoft’s existing library of over 1,800 models.

Recently, DeepSeek unveiled a complimentary AI assistant, claiming it operates on significantly less data while being much more affordable compared to established services. By Monday, this assistant surpassed the downloads of U.S. competitor ChatGPT on the Apple App Store, causing unease among investors in the tech sector.

This development aligns with Microsoft’s strategy to lessen its reliance on OpenAI, the creators of ChatGPT. The company has been focusing on incorporating both in-house and external AI models to enhance its leading AI service, Microsoft 365 Copilot, as reported by Reuters last month.

Moreover, Microsoft stated that users would soon be able to operate the R1 model locally on their Copilot+ PCs, which could help alleviate concerns related to privacy and data sharing associated with the model’s use.

DeepSeek has mentioned that it keeps user data on servers located in China, which might hinder its acceptance in the U.S. market.

In related news, Microsoft and OpenAI are investigating whether data outputs from OpenAI’s technology may have been improperly accessed by a group connected to DeepSeek, according to Bloomberg News on Tuesday.

DeepSeek’s swift rise in the AI landscape has prompted responses from competitors, with OpenAI CEO Sam Altman indicating the company would “adjust some releases.” Following this, they launched a customized version of ChatGPT for U.S. government entities on Tuesday.

Additionally, China’s Alibaba 9988.HK introduced an updated version of its Qwen 2.5 AI model on Wednesday, which is noteworthy as it coincides with the first day of the Lunar New Year.

Reporting by Deborah Sophia in Bengaluru; Editing by Maju Samuel