- Chinese President Xi Jinping endorsed open-source artificial intelligence model development at the World AI Conference in Shanghai on Friday, casting China as a champion of AI openness and equality in a speech that implicitly criticized US export controls on semiconductors and AI models; Xi’s endorsement of open-source AI is strategically calculated — China has benefited enormously from open-source AI development, particularly through Meta’s Llama models and the domestic success of DeepSeek’s open-weight models, which have allowed Chinese researchers and companies to build competitive AI capabilities without access to the most advanced Nvidia chips that US export controls have restricted.
- The speech positions China in a “digital non-aligned” geopolitical posture directed at the Global South: by championing open-source AI as a path to AI access for developing nations, Xi is offering an alternative narrative to the US framework of controlled AI access through export licensing, chip restrictions, and model access rules; countries in Africa, Southeast Asia, Latin America, and South Asia that have been unable to access US AI chips or closed frontier models are a natural audience for China’s message that open AI development serves global rather than narrow national interests — a framing that serves Chinese soft power objectives while creating friction for US AI export control policy.
- Xi’s open-source endorsement is also a response to the US’s recent tightening of AI model access restrictions: the Trump administration has imposed foreign access restrictions on certain frontier AI models (including temporarily on Anthropic’s Mythos models), and has used AI chip export controls as diplomatic leverage — most recently rewarding the UAE with expanded chip access in exchange for military support in the Iran war; Beijing is offering developing nations an alternative where access to AI tools doesn’t require geopolitical alignment with Washington, and where the most capable openly-available models happen to originate from Chinese research institutions and companies.
- The strategic irony is notable: the US has historically championed open internet and information access as democratic values, while China has built the Great Firewall; in the AI era, the US is restricting access to frontier AI through export controls and model licensing while China is advocating for open access; this role reversal in the technology access debate — whether genuine or purely rhetorical — gives China a powerful framing advantage in multilateral AI governance discussions and at institutions like the UN where developing nations have significant collective voice on technology policy.
What Happened?
Chinese President Xi Jinping spoke at the World AI Conference in Shanghai on Friday, endorsing open-source AI development and positioning China as a champion of openness and equality in artificial intelligence. Xi’s remarks implicitly criticized US efforts to protect its lead in AI chips and models through export controls and access restrictions. The speech is part of China’s broader diplomatic push to court developing nations on AI policy, offering open-source Chinese AI models as an alternative to restricted US frontier models.
Why It Matters?
Xi’s open-source AI endorsement is as much geopolitical strategy as technology policy. China has legitimately benefited from open-source AI — DeepSeek’s open-weight models and access to Meta’s Llama have let Chinese developers build at the frontier despite chip export controls — and now Beijing is weaponizing that narrative in the competition for developing-world alignment. If China successfully positions itself as the AI access provider for the Global South while the US is seen as the AI gatekeeper, it reshapes the politics of AI governance at multilateral institutions and potentially accelerates adoption of Chinese AI standards in large emerging market economies.
What’s Next?
Watch for China to follow Xi’s speech with concrete open-source AI initiatives — model releases, technical cooperation agreements, or multilateral AI governance proposals that institutionalize the “open vs. closed” framing at the UN or in BRICS contexts. Also watch the US response: the Biden administration’s AI export control framework is already under review by the Trump administration, and Xi’s speech adds political pressure on Washington to either relax restrictions (to counter China’s openness narrative) or double down on controls (accepting the geopolitical cost of being cast as the AI gatekeeper). DeepSeek’s anticipated IPO will be another signal of how China intends to commercialize its open-source AI leadership globally.
Source: The Wall Street Journal
