Beijing Moves to Lock Down Its Own AI as the Tech Cold War Escalates

China Turns the Tables on AI Openness

For the past year and a half, China's greatest weapon in the global AI race wasn't secrecy — it was generosity. Chinese tech giants freely released powerful AI models that anyone in the world could download, run, and modify. That strategy built enormous international goodwill and put real pressure on American rivals. Now, Beijing appears ready to slam the door shut.

Chinese authorities are considering tightening export controls on AI and semiconductor technologies. The move reflects Beijing's push to keep domestic AI at home and show that China, like the U.S., now sees advanced AI as a critical national asset requiring controls. It is a striking reversal — and a signal that the global AI competition has entered a more guarded, nationalistic phase.

What Beijing Is Actually Considering

Regulators led by China's Ministry of Commerce have been consulting top homegrown AI and chipmaking companies on how to stop China's advanced technologies and leading start-ups from being acquired by the west. The Commerce Ministry has also spoken with AI companies including Alibaba, ByteDance, and Zhipu on limiting the transfer of key data for the training of their models overseas, as well as allowing their model weights to be downloaded by foreign users.

Regulators have also sought views on possible restrictions that would prevent overseas chipmakers including Qualcomm and TSMC from producing advanced semiconductors based on designs developed by Chinese companies such as Huawei, Alibaba, and ByteDance. Agentic AI was suggested as one of the areas where potential restrictions could be put in place. Separately, Meta's roughly $2 billion Manus deal is also part of the debate.

The Open-Weight Dilemma

The strongest Chinese models are published as open-weight models, meaning they can be downloaded, run on a user's own servers, fine-tuned, and reused — which is precisely what has made Qwen (Alibaba), DeepSeek, and Kimi (Moonshot AI) the default choice for many developers worldwide. That strategy helped Chinese AI gain international adoption despite U.S. restrictions on advanced semiconductor exports.

For China, tighter controls would require a difficult balance. Its companies have benefited from making lower-cost models widely available, helping Chinese AI gain users and influence overseas. Broad restrictions could slow that expansion and make it harder for domestic developers to attract international customers, capital, and research partners. These proposals are still under discussion, and technology companies have apparently advised regulators that tighter measures would hinder China's ability to get ahead in the AI technology race.

A Mirror Image of Washington

The scope of the reported discussions is notable: it covers not only models already in circulation but unreleased ones, and — critically — open-weight releases of the kind exemplified by models in the Kimi and DeepSeek category, whose weights anyone can currently download. China is also tightening oversight of its AI sector by requiring key professionals at private technology companies, including Alibaba and DeepSeek, to obtain government approval before traveling overseas.

DeepSeek and Moonshot offer open-weight models that users can download, deploy on their own infrastructure, and modify for specific workloads — meanwhile, flagship models from Anthropic and OpenAI remain closed, which means Chinese companies have had an edge over rivals that they are now about to lose. The deeper irony is hard to miss: China built its AI influence by being open when America was closed. If these controls move forward, both superpowers will have chosen the same path — treating artificial intelligence not as a public good, but as a state secret worth protecting at almost any cost.