U.S. vs China: AI IP
- U.S. officials are framing artificial intelligence as a national‑security front, accusing China of stealing AI intellectual property. - The dispute now targets semiconductors, defence applications, and control of advanced systems, not just tariffs on goods. - Reporters say Washington is broadening trade policy into technology security, moving the contest beyond conventional tariffs. (economictimes.indiatimes.com)
Washington is treating artificial intelligence less like a trade issue and more like a security contest, using sanctions and export controls to keep U.S. chips, models and know-how away from China. (whitehouse.gov) The shift is visible in the rules. On January 15, 2025, the Commerce Department published an interim final rule that tightened controls on advanced computing chips and added a new control on the “model weights” of some advanced closed-weight artificial intelligence systems. (federalregister.gov) That same month, Commerce added 11 China-based entities to the Entity List, saying 10 were advancing China’s military modernization through high-performance and exascale computing, and one was supplying a sanctioned firm. On March 25, 2025, Commerce added 80 more entities across several countries, including 12 linked to advanced AI, supercomputers and high-performance AI chips for China-based end users with ties to the military-industrial complex. (bis.gov, bis.gov) The language from Washington has changed with the policy. The January 2025 rule said advanced chips and large computing clusters can enable military and intelligence uses, and the White House said on January 14, 2026 that chip action was needed to protect “economic and national security.” (federalregister.gov, whitehouse.gov) That framing reaches beyond hardware. In its July 2025 AI Action Plan, the White House called it a “national security imperative” for the United States to maintain technological dominance as rivals “race to exploit” artificial intelligence. (whitehouse.gov) The older trade case is still there underneath. The Office of the United States Trade Representative’s May 14, 2024 four-year review kept Section 301 tariffs tied to findings that China used acts and policies related to technology transfer, intellectual property and innovation that burden U.S. commerce. (ustr.gov) What is new is that the government is now trying to control the full artificial intelligence stack: the chips that train systems, the giant data centers that run them, and in some cases the model parameters that make a system work. Commerce’s January 2025 rule said those model weights can be sensitive enough to warrant export controls alongside semiconductors. (federalregister.gov) The intellectual-property piece has also moved into sanctions policy. On February 24, 2026, the State Department announced its first designations under the Protecting American Intellectual Property Act, saying theft of trade secrets tied to sensitive and emerging technologies threatens U.S. national security and economic prosperity. (state.gov) U.S. trade reporting still describes intellectual property as a barrier issue, but the 2026 National Trade Estimate and the State Department’s sanctions language place it beside export controls, cyber measures and industrial policy. That puts artificial intelligence in the same policy bucket as semiconductors, critical infrastructure and defense supply chains. (ustr.gov, state.gov) Beijing has long rejected U.S. export controls and entity listings as abuse of trade and technology rules, while Washington says the restrictions target military end uses and diversion risks. The result is that the U.S.-China fight over artificial intelligence now runs through licensing desks, sanctions lists and chip supply chains, not only tariff schedules. (bis.gov, ustr.gov)
