An AI open-source system named FlagOS 2.0 was released on Friday at the ongoing 2026 Zhongguancun Forum (ZGC Forum) Annual Conference in Beijing.

Jointly developed by 23 institutions, including leading Chinese research organizations and universities, the open-source system software platform aims to provide a unified computing infrastructure that can support multiple AI chips.

"Although many AI chips are now available, a major challenge is that different chips have different architectures and fragmented technology stacks. As a result, AI models and algorithms cannot easily run across different chips. What we need is a unified, open-source intermediate layer -- a shared technology stack -- so that once large models are released, they can be easily used across different AI hardware platforms," said Lin Yonghua, vice president and chief engineer of Beijing Academy of Artificial Intelligence (BAAI), one of the joint developers of FlagOS 2.0.

China's open-source efforts are also gaining international traction. At the forum, BAAI signed an agreement with Eclipse Foundation, which focuses on open source software collaboration and innovation, to promote joint innovation and deeper integration of global computing resources.

Eclipse Foundation is an independent, non-profit organization chartered in the European Union that acts as a steward of the Eclipse open source software development community.

"Open source is a way to collaborate on a level playing field. So I think collaboration, we heard this a lot also today in different presentations, collaboration is key for AI to build a global AI system," said Michael Plagge, Vice President of Eclipse Foundation.

This year's ZGC Forum Annual Conference opened on Wednesday in Beijing and will run through Sunday.