Export Controls: The Perennial Debate Continues

There has long been a debate in export control circles—historically, a fairly small pond—over the dilemma of overcontrolling or under-controlling exports. To oversimplify, if the United States under-controls, the bad guys get stuff we don’t want them to have. If we overcontrol, we end up kneecapping our own industries by denying them the revenue they need to invest in innovative new technologies. There is a very fine line between the two, and most administrations have walked it on a case-by-case basis by examining individual license applications carefully and by developing complicated regulations that attempt to separate relatively benign technologies from the more dangerous.

Over the years, I think the U.S. government has done a pretty good job of walking that line, recognizing that there will always be leakage. Companies make mistakes and sometimes do not undertake the necessary due diligence about their customers, and there have always been criminal elements who specialize in finding ways to get critical technologies to people we don’t want to have them.

Unfortunately, the debate regarding controls, both in Congress and within the executive branch, has not always been enlightened and thoughtful. The main reason involves a descent into the world of “could.” Those who favor tighter controls—the over-controllers—argue that failure to control a specific item “could” lead to a national security problem. The under-controllers argue that the probability of any of the alleged parade of horribles actually happening is near zero. But it is not zero, and that is why they often lose the argument. They cannot say with 100 percent certainty that bad things cannot happen if a particular item is exported.

That debate, which I experienced during my time in government and have observed ever since, took a new turn recently with the agreement to sell millions of high-end Nvidia and Advanced Micro Devices AI chips to entities in Saudi Arabia and the United Arab Emirates (UAE), most of which are U.S. companies.

The impetus for the deals appears to have come from David Sacks, who is the White House AI and crypto czar. His argument is a variation of the one explained above and is remarkably similar to the one used during the initial debate over exports of encryption technology in the late 1990s. The argument then was that if countries were determined to buy encryption, it would be better for our economy and our security if they bought American. Sacks’ argument now seems to be that the United States should encourage the world to use U.S. chips because if we don’t, others will turn to China, and we will lose our lead in AI technology. At an event in Saudi Arabia during Trump’s visit, Sacks said, “We need our friends, like the kingdom of Saudi Arabia and other strategic partners and allies, to want to build on our tech.”

In my view, this is the right policy, but the administration is applying it to the wrong countries. This parallels the debate over “friend-shoring”—building supply chains in reliable, friendly countries in order to maintain resilience. A similar term we have used at CSIS is “trusted partner.” The underlying challenge—often left unstated—is deciding who is a friend and who is trusted. Sometimes, it’s obvious. NATO allies, for example, generally fall into both categories—though even among them, questions can still arise. For example, is Turkey a friend? Outside of formal allies, the situation is even murkier. India seems to be a new friend. Does that make it a trusted partner? How long might that friendship last?

In the export control context, is Saudi Arabia a friend? Can it be trusted? What about the UAE? If you talk to people in the export control enforcement business, they will tell you “no.” Both are notorious sieves, and there is very little confidence in the enforcement community that technology exports to them will stay there. More likely, they will end up in Russia or China after passing through many intermediaries. Both Saudi Arabia and the UAE seek to maintain good relations with the United States and with China at a time when both want them to choose between them. That may very well be a rational policy on their part, but it is not one that creates confidence in the United States that they will be reliable recipients of U.S. technology.

That reality takes us back to the initial dilemma of overcontrolling or under-controlling. Sacks is right that the United States should be pursuing policies to ensure U.S. tech leadership globally over the long term, and that includes trying to dominate the market for cutting-edge technologies so that we don’t cede leadership to China. But he is naive in thinking that Saudi Arabia and the UAE are trusted partners and will be reliable end users of U.S. semiconductors. Protecting our interests still requires case-by-case and country-by-country analysis, and the administration should not lose sight of that in its search for headline-making deals.

William A. Reinsch holds the Scholl Chair in International Business at the Center for Strategic and International Studies in Washington, D.C.