China proposes export controls on AI models and chip technologies

By Splitfeed EditorialLast updated Oct 19, 2026, 04:00 UTC

In mid-October 2026, China's government proposed revisions to its export control list that would cover 23 AI-related technologies, a category that includes AI models, chip technologies, and solar technologies. The stated rationale, as reported in mid-October, centers on securing critical infrastructure and government systems. Separately, Taiwan announced around the same period that it would impose strict controls on AI chip exports directed at China. Both moves landed in the same week that official figures showed China's economy growing at a stronger pace, with AI-driven exports credited as a contributing factor, even as weak domestic demand continued to limit the overall picture.

The timing matters because it places these proposals inside an accelerating cycle of mutual technology restrictions between China, Taiwan, and the United States. China's list of 23 controlled technologies is notable for explicitly including AI models, not just the hardware that runs them, which would mark a meaningful expansion of what governments treat as a controllable export. If adopted, such controls could complicate how Chinese AI research and commercial products move across borders, and Taiwan's parallel chip-export restrictions suggest the pressure is being applied from multiple directions simultaneously. The fact that China's exports are already receiving an AI-related lift raises the question of whether these controls are primarily defensive, aimed at protecting proprietary systems, or strategic, aimed at managing what competitors can access.

Because no reactions from the Pro-AI, Anti-AI, or Middle Ground camps have been published yet, the positions below reflect what each camp would typically argue about a story of this kind rather than anything anyone has said. The Pro-AI camp would likely argue that export controls risk fragmenting global AI development, slowing progress that benefits everyone, and that competition between nations is better handled through investment and openness than through restriction. The Anti-AI camp would more probably welcome controls as a rare case of governments asserting meaningful oversight over powerful technologies, though it might question whether self-interested national security framing is the right basis for that oversight. The Middle Ground camp would likely focus on the difficulty of drawing workable boundaries around AI models as exports, noting that software controls are historically hard to enforce and that poorly designed rules could harm researchers and businesses without achieving their security goals.

The key things to watch are whether China finalizes the proposed revisions and publishes the precise definitions it will use for controlled AI models, how Taiwan implements its chip-export restrictions in practice, and whether the United States responds with further measures of its own, any of which would clarify whether this week's announcements represent a genuine escalation or an opening negotiating position.