export controls on AI chips
In plain wordsGovernment rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI.
Restrictions on exporting advanced AI acceleratorsAI accelerator, n.: Any chip built or tuned to speed up AI — graphics chips, Google's TPUs, phone 'neural engines' and many startup designs. Such chips do the heavy math of training and running models., the semiconductor manufacturing equipment needed to make them and related technology, used chiefly by the United States, with allied cooperation from the Netherlands and Japan, to slow China's military and frontier-AI development. US rules, administered by the Commerce Department's Bureau of Industry and Security (BIS), began on October 7, 2022 with controls on chips above performance and interconnect-bandwidth thresholds; an October 2023 update switched to a 'total processing performance' metric, closing gaps exploited by China-specific chips such as Nvidia's A800 and H800, and later rules added high-bandwidth memoryhigh-bandwidth memory, n.: Very fast memory made by stacking memory chips on top of each other and placing them right next to the processor, so AI chips can be fed data quickly enough. (December 2024) and more equipment. The AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy. of January 2025 tried to extend controls worldwide but was rescinded in May 2025. Policy then turned transactional: after requiring licenses for Nvidia's H20 in April 2025, the administration let sales resume in August in exchange for 15% of the revenue, and from mid-January 2026 Commerce reviewed licenses for exports of H200- and MI325X-class chips to China case by case, subject to conditions including volume caps and testing in the United States, with 25% of the value going to the US government. The debate pits national security against market share and the risk of accelerating China's domestic chip industry.
Export controls on AI chips mean the lab must train on older accelerators or rent compute abroad.
Origin
Rooted in Cold War export-control regimes (COCOM, later the Wassenaar Arrangement) and US measures against Huawei from 2019; the BIS rules of October 7, 2022 were the first controls aimed squarely at AI computing.
Usage note
Enforcement contends with smuggling, shell companies and remote cloud access, for which chip location verification and compute KYCknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account. are proposed fixes. China has answered with its own controls on critical minerals and a push for domestic accelerators such as Huawei's Ascend. Export-control powers also reach models: see export control directiveexport control directive, n.: A US government order, used against AI models in June 2026, telling a company to stop foreigners from using some of its AI models for national-security reasons..
Related
- AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy.
- compute governancecompute governance, n.: Using control over the computer chips and data centers that AI depends on — who can buy them, where they go, how they are used — as a way to oversee powerful AI.
- on-chip governance mechanismson-chip governance mechanisms, n.: Features built into AI chips that could check or enforce rules about how they are used — for example, proving which country a chip is in, or refusing to run without a valid license.
- know-your-customer for computeknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account.
- export control directiveexport control directive, n.: A US government order, used against AI models in June 2026, telling a company to stop foreigners from using some of its AI models for national-security reasons.
- CHIPS and Science ActCHIPS and Science Act, prop. n.: A 2022 US law that put tens of billions of dollars into building computer-chip factories and chip research in the United States, to reduce dependence on chips made in Asia.
- NvidiaNvidia, prop. n.: An American chip company whose graphics processors became the main hardware for training and running AI, making it one of the most valuable companies in the world.
- AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety.
Sources
- Bureau of Industry and Security (2022). Commerce Implements New Export Controls on Advanced Computing and Semiconductor Manufacturing Items to the People's Republic of China (PRC). Press release, October 7, 2022.
- Covington & Burling (2026). U.S. Commerce Department Revises License Review Policy for Exports of Certain Advanced Computing Commodities to China and Macau. January 2026. cov.com/en/news-and-insights/insights/2026/01/us-commerce-department-revises-license-review-policy-for-exports-of-certain-advanced-computing-commodities-to-china-and-macau
Defined in law or standards
US Export Administration Regulations (15 C.F.R. Parts 730–774), ECCNs 3A090 and 4A090 (advanced computing items)
Mentioned in
AI acceleratorAI accelerator, n.: Any chip built or tuned to speed up AI — graphics chips, Google's TPUs, phone 'neural engines' and many startup designs. Such chips do the heavy math of training and running models., AI arms raceAI arms race, n.: A competition among countries or companies to build the most powerful AI first, in which fear of falling behind pushes everyone to move faster and cut corners on safety., America's AI Action PlanAmerica's AI Action Plan, prop. n.: The Trump administration's 2025 strategy for winning a global AI race: cutting regulation, building data centers and power plants quickly, and exporting American AI technology to allies., DeepSeekDeepSeek, prop. n.: A Chinese artificial intelligence company whose free, cheaply trained models startled the world in January 2025 by nearly matching the best American systems., Jensen HuangJensen Huang, prop. n.: The Taiwanese-American co-founder and chief executive of Nvidia, whose graphics chips became the main engine for training and running modern AI., NvidiaNvidia, prop. n.: An American chip company whose graphics processors became the main hardware for training and running AI, making it one of the most valuable companies in the world., ASMLASML, prop. n.: A Dutch maker of the giant machines that print circuit patterns onto chips. It alone can build the extreme-ultraviolet models on which today's most advanced chips, including AI chips, depend., CHIPS and Science ActCHIPS and Science Act, prop. n.: A 2022 US law that put tens of billions of dollars into building computer-chip factories and chip research in the United States, to reduce dependence on chips made in Asia., compute governancecompute governance, n.: Using control over the computer chips and data centers that AI depends on — who can buy them, where they go, how they are used — as a way to oversee powerful AI., DeepSeek momentDeepSeek moment, n.: The shock of January 2025, when a little-known Chinese company released a free AI model nearly as good as the best American ones at a fraction of the reported cost, briefly wiping hundreds of billions off chip stocks., dual-usedual-use, adj.: Describes technology or knowledge that can be used both to help and to harm: the same AI that finds software bugs for defenders can find them for attackers., H100H100, prop. n.: Nvidia's flagship AI chip of 2022–2024, so scarce and sought-after during the generative-AI boom that it became a unit of measure for AI computing power., HuaweiHuawei, prop. n.: A Chinese technology giant, known for telecom equipment and phones, that designs China's leading homegrown AI chips despite US trade restrictions., Liang WenfengLiang Wenfeng, prop. n.: A Chinese entrepreneur who founded DeepSeek, whose low-cost, openly released AI models shocked global markets in January 2025., process nodeprocess node, n.: The label for a generation of chip-making technology, like '3 nanometer'. Smaller numbers mean newer, denser chips, but the number no longer measures any actual part of the chip., semiconductor foundrysemiconductor foundry, n.: A factory company that manufactures chips designed by other companies. Taiwan's TSMC, the largest, makes nearly all of the world's most advanced AI chips., sovereign AIsovereign AI, n.: A country's ability to build and run AI on its own terms, so that the computers, models and know-how it relies on cannot be withheld or dictated by foreign companies or governments., TSMCTSMC, prop. n.: The Taiwanese company that manufactures most of the world's advanced computer chips, including nearly all of the specialized chips used to train and run AI., AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy., Center for Security and Emerging TechnologyCenter for Security and Emerging Technology, prop. n.: A Georgetown University think tank in Washington, D.C., whose researchers advise the US government on AI and national security, paying special attention to China and to the chip supply chain., export control directiveexport control directive, n.: A US government order, used against AI models in June 2026, telling a company to stop foreigners from using some of its AI models for national-security reasons., extreme ultraviolet lithographyextreme ultraviolet lithography, n.: The chip-printing technique used for the most advanced chips, which draws circuit patterns with ultra-short-wavelength light. A single Dutch company, ASML, makes all the machines, each costing well over $100 million., know-your-customer for computeknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account., Mutual Assured AI MalfunctionMutual Assured AI Malfunction, n.: A proposed deterrence strategy in which any country racing for overwhelming AI dominance would expect rivals to sabotage its AI projects, so none tries, much as with nuclear deterrence.