# U.S.-China Chip Controls Could Enforce AI Pause Lasting at Least 10 Years

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/29705
Published: 2026-10-11T04:45:12.000Z
Updated: 2026-10-11T04:45:12.000Z
Section: Technology

> A proposed system would halt or sharply reduce training-chip production while allowing whitelisted services such as ChatGPT and Claude to continue operating.

A proposed global pause on training more powerful artificial intelligence models could be enforced through advanced-chip controls if U.S. and Chinese leaders wanted to halt or sharply reduce training-chip production, an analysis found.

The pause would last at least 10 years and would allow existing AI services to continue if they were placed on a whitelist. New frontier-model training would be prohibited, while chipmakers would shift toward hardware designed only to run existing models.

A roughly 200-page analysis released Oct. 9 by 26 scholars, led by Berkeley professors Will Fithian and Wesley Holliday, examined whether governments could stop frontier-model training while keeping approved services operational.

The proposed system would also require countries to inventory training chips already in circulation. Existing hardware could initially be used for inference, the process of running an existing model, under verification controls. Governments could later phase out those chips through buyback programs and exchanges for newer inference-specific hardware.

The analysis estimated that global training capacity would reach the equivalent of 47 million Nvidia H100 chips by the end of 2026, with about 2.9 million additional chips shipping each month. Replacing the global stock of training hardware would take about 16 months at that pace, while the broader transition could be completed within five years.

Small-scale secret training would be difficult to detect, but developing a model beyond the current frontier would require hundreds of thousands to millions of chips operating for months. A country secretly seeking a strategically significant model within five years could need about 6.4 million chips.

A country openly abandoning the proposed arrangement and pursuing a decisive model within two years could need about 180 million chips, nearly four times the estimated global stock.

The analysis identified two vulnerabilities: the United States could retain enough training hardware used for transitional inference to restart training, while undeclared computing capacity in China could support covert development.

Supply-chain controls would be central. ASML has a monopoly on manufacturing the extreme ultraviolet lithography machines needed for the most advanced chips. Taiwan Semiconductor Manufacturing Co. operates leading-edge foundries and packaging facilities, while Nvidia chips account for more than 60% of global AI computing capacity.

The political conditions for such an arrangement are not currently in place. Leaders from more than 20 countries responded to a Sept. 21 appeal launched by Finnish President Stubb and Norwegian Prime Minister Støre to control frontier AI models, but the United States and China did not participate.

The broader economic impact could be limited because most businesses do not use frontier models. Even a $6.5 trillion decline in the market value of the seven largest U.S. technology companies to early-2025 levels would reduce U.S. consumer spending by about $195 billion, or 0.6% of 2026 gross domestic product.

Nvidia would face a particularly difficult adjustment because of its heavy investment in training-class chips. Shifting toward inference-specific hardware could be difficult and expensive, while the pause could strengthen dominant companies and leave AI models with increasingly outdated knowledge.