Mismatch of Strategy and Budgets in AI Chip Export Controls

Remote Visualization

This commentary is part of a report from the CSIS Economic Security and Technology Department, titled Staying Ahead in the Global Technology Race. The report features a set of essays outlining key issues on economic security for the next administration, including global technology competition, industrialization policies, economic partnerships, and global governance.

Two dates from 2022 are likely to echo in geopolitical history. The first, Russia’s full-scale invasion of Ukraine on February 24, hardly needs further explanation. The second is one that many Americans may not recognize. On October 7, 2022, the U.S. Department of Commerce issued new export control regulations that placed a de facto ban on U.S. sales to China of the most advanced computer chip hardware that powers modern artificial intelligence (AI) models.

The United States and China agree that leadership in AI technology is critical to the future of military power. For years, Chinese government and military procurement records openly advertised the desire for U.S. chips to power Chinese AI surveillance systems and new AI military supercomputing facilities. Since more than 90 percent of AI chips used in Chinese data centers are designed by U.S. semiconductor companies and are therefore subject to U.S. export controls, loss of access to the U.S. chip market could put China’s entire future as an AI superpower in jeopardy.

Grand historical turning points rarely take the form of long bureaucratic documents, but the October 7 export controls were one of those rare times. Ten days after the new regulations came out, Secretary of State Antony Blinken said: “We are at an inflection point. The post–Cold War world has come to an end, and there is an intense competition underway to shape what comes next. And at the heart of that competition is technology.”

Blinken is right. Even though the October 7 export controls were in many ways narrowly targeted on only the most advanced AI chips and chipmaking tools, as a whole, the policy marked a major reversal of over 25 years of trade and technology policy toward China in at least three ways.

First, the controls were targeted at multiple chokepoints across the semiconductor supply chain, blocking sales not only of the advanced AI chips being used by the Chinese military but also the advanced software and equipment required to make them. The United States is trying to ensure that China cannot replace what the United States is no longer willing to sell.

Second, the export controls apply on a geographic basis for China as a whole, not just to the Chinese military. That is a response to China’s strategy of military-civil fusion, which has worked to deepen and obscure the linkages between China’s commercial technology companies and China’s military.

Third, previous U.S. export controls were designed to allow China to progress technologically but to restrict the pace so that the United States and its allies retained a durable lead. The new policy, by contrast, in some cases aims to actively degrade China’s technological capabilities. Without access to the United States’ enabling technology, many leading Chinese semiconductor firms have been set back years.

It took a long time for the United States to get here. After decades of ratcheting Chinese government provocations, the Biden administration basically said, “enough is enough.”

This is not a policy of decoupling (yet), but it is proof of the United States’ unwillingness to remain tightly coupled to the Chinese technology sector under previous conditions. Subsequent policies, such as the Treasury Department’s outbound investment restrictions on China’s AI and semiconductor industries, hint at the United States’ desire for more comprehensive economic security and technology.

But there is a critical gap between the strategic importance and sophistication of the policy’s design and the resources that the government is allocating to enforce it.

The Bureau of Industry and Security (BIS) at the U.S. Department of Commerce is the agency charged with enforcing export controls, not just on semiconductors bound for China but for all U.S. dual-use technology exports that might end up in Russia, Iran, North Korea, or other restricted destinations. To implement its work overseeing trillions of dollars in economic activity and policing smuggling operations worldwide, BIS has fewer than 600 employees and a relatively paltry budget of just under $200 million. Semiconductors are just one technology category out of hundreds that this organization is responsible for enforcing.

Reporting by The Information found at least eight Chinese AI chip-smuggling networks, with each engaging in transactions valued at more than $100 million. China is betting that its network of smugglers and shell companies can find the leaks in the BIS export control enforcement barrier. As long as Congress continues to neglect BIS by providing grossly inadequate resources compared to the size and importance of its mission, China has a reasonable expectation of success. BIS needs not only more money, but also more skilled staff, more enforcement agents, and better enabling technology, especially for data analysis.

Moreover, the Department of Commerce needs more help from the rest of the government, in particular the U.S. intelligence community. Declassified Central Intelligence Agency documents show that the intelligence community was deeply involved in assisting export control enforcement during the Cold War and delivered solid results by doing so. These are capabilities and priorities that have significantly atrophied in the post–Cold War era but urgently need to be restored.

Regardless of who wins the November 2024 election, export control represents a great deal of unfinished business for the next presidential administration to take on.

Finally, the United States cannot do this alone. U.S. allies need to take a good look at their own export controls and broader economic security toolboxes. There are some innovative economic security policy experiments going on in places like Taiwan, South Korea, and Japan. Allies need to share information on best practices, align approaches, and devote appropriate resources to have a reasonable chance of success.

Gregory C. Allen is the director of the Wadhwani Center for AI and Advanced Technologies at the Center for Strategic and International Studies in Washington, D.C.