At the World Artificial Intelligence Conference in Shanghai, President Xi Jinping announced China’s push for low‑cost, open‑source AI, framing it as a global public good contrasting with the US model.

China is building a self‑sustaining AI ecosystem, offering it abroad and embedding its standards and governance. US export controls have inadvertently become a catalyst, forcing companies like Huawei to develop open‑source systems such as Harmony OS and Mind Spore. The result is a wave of open‑weight models—publicly available weights but private training data—making AI cheaper and adaptable for emerging markets.

Domestic demand is weak, so exporting lower‑cost models spreads China’s technology quickly and eases reliance on advanced chips. However, releasing weights raises governance challenges: models can be modified, redistributed, and stripped of safeguards, prompting China to explore “selective openness” and new safety frameworks.

China’s ambition is to broaden its strategic foothold, diversify geopolitical risk and reshape the global tech system in its favour, even as its governance structures remain fluid and evolving.