British historian Paul Kennedy wrote in “The Rise and Fall of the Great Powers” that great powers do not decline because they fail to recognize problems, but because they apply old solutions to new problems when those solutions are no longer effective. This broad statement aptly describes US technology transfer policy, which is still derived from the bifurcated world of the late 20th century when the US had an unassailable lead in technology. But that era is over. China is not the sclerotic Soviet Union, the world is not sharply and impermeably divided between East and West, and technology is no longer the product of defense-centric, government-led innovation of the 1980s. Those were the conditions export controls were designed for—and they no longer exist. What exists now is a highly interconnected and interdependent world driven by digital technologies.
The lessons of Afghanistan and Iraq are that having the best technology does not produce victory. While it cannot be said the US was defeated in its Middle Eastern wars (although some of the domestic turmoil we see today derives from these failures), the US was unable to achieve its objectives. Technology is one factor among many for military success. Having the “lead” in AI in a contest between near peers (or even less-advanced but determined opponents) will not ensure success. The variables that ultimately determine success in conflict are political will and strategy. The first (derived from Clausewitz) is that only when one side no longer has the will to contest the outcome of a conflict can it be defeated. This is unlikely to occur with China. The second variable, strategy, has been the great weakness of American military efforts. The strategic weakness of the policies issued in the final moments of the last Administration are that they are focused on the wrong goals. Races are not won by selling, not denying.
When a senior US official announced in 2003 that the US would bring democracy and a modern society to Afghanistan, this failed to recognize the will of Afghans to resist change brought by foreign occupiers. When another senior official said that the conquest of Iraq was a first step in bringing stability to the Middle East, it was another error—in this case, failing to understand the requirements of such a bold undertaking as occupying Iraq. American goals were frankly unrealistic and unattainable, and the outcome of the conflict is that the US has suffered damage domestically and internationally as a consequence.
Effective strategy requires identifying goals and the actions to attain them. American strategy and goals vis-à-vis China remain inchoate. If the goal is to prevent China, the world’s second-largest economy with important and improving scientific strengths, from modernizing its military, we will not succeed. China has many problems of its own making, but restrictions on dual-use technology transfer are, with a few exceptions, largely irrelevant to China’s power.
In contrast, if the objective is to prevent China from continuing to use illicit and predatory practices to achieve global economic and political preeminence (a goal enunciated by China’s leader), this is an attainable goal, but not through military means. An over-militarized approach to technology transfers creates more risk than benefit, since the likely result of seeking to deny China access to technologies is to incentivize it to create substitutes.
We also need to ask what role and contribution technology transfer restrictions, such as export controls, play. A case can be made that their contribution in the larger contest is marginal. To qualify this, restrictions on the sale of weapons, munitions, and military equipment remain essential. The US does not want to arm its opponents and can reasonably expect its treaty allies to avoid this as well. Restrictions on technologies related to the proliferation of weapons of mass destruction similarly remain essential. This is a limited set of goods identified by the non-proliferation regimes, and restrictions have limited commercial effect.
That leaves dual-use goods, industrial goods, components, and manufacturing equipment (including software) that can contribute to the construction of a more advanced military force. This is where most of the friction over technology transfer policy exists today, and it is the category whose contributions are open to question, with the chief questions being how much they retard China, what costs they impose on Western nations, and what is their value if not linked to a coherent and achievable strategy.
The discussion of current export controls overlooks their origins as a relatively minor Cold War policy tool, primarily intended to prevent America’s allies from selling advanced industrial goods to the Soviet Bloc. Initially, controls on dual-use goods applied only to the Soviet Union and its allies. These Cold War-era controls, despite their limited scope, remain the foundation of the current system. Understanding this history helps explain the weaknesses of present-day export controls and has significant implications for their potential success.
A global export regime emerged between 1991 and 1993 as part of a broader American foreign policy initiative. This development was prompted by the collapse of the Soviet Union, growing non-proliferation concerns, and the first Persian Gulf War. The decision to create global export controls in 1993 received direct Presidential support and was closely linked to a major diplomatic effort to integrate Russia and Warsaw Pact countries into the Western order.
European support for export controls was selective: they backed restrictions on weapons of mass destruction (WMD), and on exports to Iran, Syria, and Iraq, but notably excluded China. The current approach lacks the direct Presidential engagement that characterized the Bush and Clinton administrations, who personally worked with their counterparts. Moreover, the global economic landscape has transformed dramatically. Trade with the Warsaw Pact and China was minimal in 1990, before the tidal wave of globalization triggered by the Cold War’s end and the emergence of technological capabilities in countries like India, Turkey, and Brazil.
The overarching foreign policy goal was to incorporate the Soviet Union and Warsaw Pact members into the Western order. A secondary concern was developing a new security architecture to address the proliferation of WMD and conventional weapons. The tacit Cold War understanding between NATO and the Warsaw Pact regarding arms transfers had already begun to unravel, particularly evident in the case of Iraq, which had assembled the world’s fourth-largest military through arms purchases from both Western and Soviet sources.
Following the Gulf War, the Bush Administration, alongside partners in the UK and France, initiated two parallel efforts: one in the G7 focusing on dual-use goods and another in the P-5 addressing conventional arms transfers. The Clinton Administration merged these efforts, creating a new regime initially centered on conventional arms. Almost as an afterthought, this new regime absorbed CoCom (the Coordinating Committee), a NATO body that had collectively managed dual-use technology transfer to the Soviet Union. The Soviet-focused regime would be replaced by a global arrangement in which Russia and other Warsaw Pact members would be equal participants.
Confusion characterized this transition. Russians, still reeling from the USSR’s rapid collapse, often misunderstood the scope of changes they were accepting. Their perception was clouded by their experience with the Warsaw Pact’s COMECON, which had coordinated trade and production among member countries. Some Russian officials assumed CoCom operated similarly, expecting market shares to be allocated rather than competed for openly.
Western allies were not united in their approach. CoCom had been a consensus-based organization where all members had to agree on proposed exports, effectively giving the United States a veto—a power the U.S. (particularly the Department of Defense) was eager to maintain, while Europeans sought to eliminate. This dynamic may have contributed to Russian misunderstandings. Some European allies even proposed that export controls on dual-use technology were unnecessary, believing they were entering the “end of history” with diminished interstate conflict.
The mid-1990s efforts resulted in the revitalization of WMD regimes and the creation of the Wassenaar Arrangement. Despite gaining renewed relevance in the tech competition with China, the Wassenaar Arrangement remains the least powerful of the non-proliferation regimes. This weakness stems partly from the less compelling threat of dual-use manufacturing equipment compared to missiles or chemical and biological weapons, and partly from ongoing disputes about national discretion in export approvals, control list modernization, and the treatment of China.
The rationale for restrictions on dual-use goods began to fray in the 1990s. With the end of the Cold War, a few major US allies even called for dual-use controls to simply be eliminated. The US opposed this, in part because it hoped to prevent technology transfer to China and in part because creating a new dual-use regime provided an opportunity to engage Russia and other Warsaw Pact states in a common international security program (some of the details of which have not yet been made public). There was no support among allies for the US effort to restrict dual-use transfers to China (arms transfers by allies had been suspended after the Tiananmen massacre), but the second goal had some success, with even the Russians cooperating with constraints on exports for almost two decades.
Since the mid-1990s, the technological and political context for restrictions has changed fundamentally. There has been a diffusion of scientific and technological skills that undermines dual-use controls. When access to a technology is blocked, countries either find substitutes or build competing industries. This was the experience of American export controls on machine tools, satellites, high-performance computers, night vision goggles, and encryption.
In some cases, the failure to relax American restrictions permanently damaged the global competitiveness of American companies. In other cases, countries with access to science and research built competing industries. For example, the European Union developed satellite components to replace those captured by US regulation (which required re-export approval from the US for EU satellites that had American components), advertising them as “ITAR-Free” to highlight to the market their independence from onerous American extraterritorial regulations. A more recent example is the decision not to sell UAVs to Turkey, which has since developed a self-made world-class drone industry that produces high-quality platforms at much lower costs.
The political context for technology transfers, originally favorable to expanded trade with China, has also dramatically changed since 1990. At that time, Western countries, including the US, were eager to engage commercially with China and flooded the country with money and technology to take advantage of its market and labor. These policies were driven by the traditional hope that commerce reduces the chances of conflict and would transform China into a market democracy. Some in the business community still profess to nurture this hope, but the arrival of Xi Jinping ended it conclusively.
Furthermore, China’s predatory economic practices, massive espionage, human rights violations, and expansionist claims—present before Xi but greatly expanded under his rule—changed global attitudes about the likelihood of China becoming a responsible international actor. It became clear that China would engage on the international stage only on its own terms, which are centered on preserving and expanding the dominance of the Chinese Communist Party.
It is well past time to reconsider the trade and technology policies pursued in the 1980s. This project will look at how to reframe US tech transfer policy for the current century, not the last. It will focus on policy and regulation for exports of advanced chips for data centers and recommend ways to make it more successful by setting proactive goals. Previous efforts, focused on denying China technology, risk doing more damage than good. Efforts to block access to advanced GPU chips valuable for AI services saw the emergence of DeepSeek, a cheaper and attractive Chinese AI model that appeared two years after ChatGPT. China is not going away, nor is it giving up. The US goal must become enabling its companies—which still have superior AI and data-center technologies—to outcompete China in the global market.