China Expands Open-Source Ecosystem with AI, Robotics

China has significantly expanded its open-source ecosystem, adding seven new projects covering artificial intelligence, embodied intelligence, electric power, manufacturing, chips, and instruction sets at the 2026 OpenAtom Open Source Ecosystem Conference held June 25-26 in Beijing. The expansion, reported by CCTV News, signals China’s accelerating commitment to open-source development as a national strategic priority, with the promotion of open-source system construction formally written into the 15th Five-Year Plan (2026-2030).

Context: A Strategic National Priority

China has been systematically building its open-source ecosystem as a top-level government priority. The OpenAtom Foundation (开放原子开源基金会), established to steward open-source projects, serves as the primary institutional vehicle for this strategy. According to the Xinhua News Agency, China’s open-source ecosystem is now transitioning from scale expansion to a new phase of high-quality development.

As of 2025, China boasted over 4.25 million active open-source projects and more than 2.63 million active open-source developers. The country has also emerged as a global leader in open-source AI: Chinese-developed models accounted for 41% of all large model downloads on major international platforms in the past year, and Chinese open-source models’ global downloads surpassed the United States for the first time in 2025, ranking first worldwide.

Key Developments at the OpenAtom Conference

The conference, themed “Open Source Empowers Industry, Ecology Builds the Future,” featured several major announcements:

Seven New Projects Donated to OpenAtom Foundation: The new projects span critical technology domains including the Dianhong IoT Operating System, LoongArch Binary Translator, and Longxi (OpenAnolis), covering AI, embodied intelligence, electric power, industrial systems, operating systems, and instruction sets. The Magnolia License Family, comprising six open-source licenses suitable for different scenarios, was also donated.

Embodied Intelligence Takes Center Stage: A humanoid robot demonstration at the Beijing event showcased over 20 different actions, many developed by global users within the open-source community. Wang Song, a company representative in robotics, stated that his organization has open-sourced approximately 60,000 minutes of real-world robot dataset, with nearly 16,000 developers in their community. “This allows ordinary developers or even novice users to engage in robot or embodied intelligence application development with a low barrier to entry,” Wang said.

New Communities Launched: The OpenAtom LoongArch Community was officially launched as China’s first comprehensive open-source community for autonomous instruction sets, with a focus on RISC-V architecture. The OpenAtom Yihong Community was also established, dedicated to IoT systems for instrumentation and control systems.

Government and Industry Voices

Xiong Jijun (熊继军), Vice Minister of the Ministry of Industry and Information Technology (MIIT), outlined the government’s vision: “Enhance open-source supply capabilities, increase the traction of open-source development in areas such as basic software, industrial software, and artificial intelligence. Support the application and implementation of excellent open-source products, and support top global open-source projects and developers to come to China for development.”

Li Bo (李博), Deputy Secretary-General of the OpenAtom Foundation, emphasized the strategic importance of the new projects: “In 2026, in terms of frontier technology layout, projects including artificial intelligence, embodied intelligence, and instruction sets have been introduced to seize the initiative in new architectures such as large models and RISC-V, using an open foundation to break technological monopolies.”

Analysis: Dual Strategic Motivations

China’s open-source expansion serves dual purposes. First, it fosters domestic innovation while reducing dependence on foreign proprietary technologies — a particularly significant goal given ongoing US-China technology tensions and export controls on advanced chips. The focus on RISC-V and autonomous instruction sets through the LoongArch community reflects this strategic autonomy drive.

Second, by open-sourcing AI models and robotics datasets, China positions itself as a global technology contributor, building soft power and developer ecosystems that extend beyond its borders. As noted in a People’s Daily opinion piece published via CCTV, “Chinese AI companies’ choice of open-source technology path has driven the explosive development of global open-source models,” according to Bill Ready, CEO of Pinterest.

The inclusion of electric power, manufacturing, and industrial operating systems in the new projects also indicates a focus on real economy applications, aligning with China’s broader “new quality productive forces” (新质生产力) policy framework. With over 6,000 AI companies operating in China and daily token calls exceeding 140 trillion, the country’s AI industry scale provides a substantial foundation for continued open-source growth.

What to Watch

As China’s open-source ecosystem matures, several key questions remain. The OpenAtom Foundation’s governance model for incubating new projects will be critical to their success. MIIT Vice Minister Xiong’s call to “improve open-source security governance rules” suggests awareness of potential risks in widespread open-source adoption, particularly in critical infrastructure. Additionally, the emphasis on exploring open-source business model innovation indicates recognition that sustainable ecosystems require viable commercial pathways.

With open-source development now embedded in China’s Five-Year Plan framework, the trajectory is clear: China is betting big on open source as both a technological strategy and a geopolitical instrument, aiming to shape the future of global technology development on its own terms.