Countries must move beyond viewing AI as a competition
The U.S. and China in the AI Landscape
On December 9th, U.S. President Donald Trump announced that the United States would allow the export of Nvidia’s H200 processors to China, but with a 25% fee on all sales. This decision has sparked significant debate within the American political landscape, with critics like Senator Elizabeth Warren accusing Trump of compromising national security by "selling out" the country's interests.
The global artificial intelligence (AI) space is often framed through a zero-sum lens, where one nation's gain is perceived as another's loss. Companies like Anthropic emphasize AI safety at home, while their co-founder and CEO, Dario Amodei, highlights an arms race abroad, suggesting that export controls are necessary to slow down China's AI development and secure a U.S. victory in the AI race. Similarly, author Chris Miller, in his book Chip War, argues that U.S. chip export controls have significantly hindered China's chipmaking capabilities.
Trump himself has repeatedly emphasized that the U.S. started the AI race and will win it. However, this framing overlooks the complexities of the geopolitical AI ecosystem. From a rational choice perspective, the term "AI race" may be misleading. A two-player race typically involves a rivalrous resource that cannot be enjoyed by both parties, which is not the case in the AI domain.
Understanding the AI Ecosystem
The geopolitical AI ecosystem differs from traditional competitive scenarios. The use of AI models is excludable, as seen when Sam Altman decided to exclude Chinese users from OpenAI’s GPT. However, such use is not strictly rivalrous, as open-source models like those from DeepSeek can be accessed and used by anyone. While the implementation of models may involve rivalrous elements due to energy and data costs, these were not the primary concerns behind Altman’s decision. Instead, he excluded Chinese users based on the belief that the U.S. should not cooperate with China.
Some argue that selling chips to China could embolden Beijing and harm U.S. interests. Yet this view neglects the benefits that U.S. middle-class households gain from access to advanced electronics at lower prices. Additionally, the global dependence on American tech provides leverage that should not be overlooked.
Economists refer to situations involving non-rivalrous but excludable resources as a "stag hunt," inspired by philosopher Jean-Jacques Rousseau’s A Discourse on Inequality. In this scenario, hunters can choose to collaborate to catch a large prey (the stag) or work alone to catch a small one (the rabbit). Cooperation leads to better outcomes, yet mistrust can lead to harmful mistakes, such as overestimating threats or reckless AI deployment in conflicts.
Cooperation Over Competition
Global AI competition resembles a stag hunt rather than a race. Whether in policy, governance, or trade, cooperation between countries can yield greater benefits than working in isolation. A breakdown in communication breeds mistrust, which could lead to escalatory spirals or reckless AI deployment in conflicts. The "stag" in the U.S.-China AI game lies in mutual prevention of such mistakes and the gains from mutually advantageous commercial development of AI for the wider public.
Common challenges exist that China, the U.S., and the world must address, including AI manipulation, deception, and coercion, as well as labor displacement caused by AI implementation. These issues require trust, transparency, and cooperation, rather than erratic politicization. Moving from hunting the rabbit to hunting the stag necessitates effective multilateral AI governance institutions, including dispute resolution mechanisms.
Building Bridges in a Multi-Polar World
Policymakers must seek to cultivate effective multilateral AI governance institutions, including establishing and monitoring dispute resolution mechanisms. Bargaining capital also arises through unconventional alignments of medium-size powers, each with their distinct niches.
For example, energy-rich Saudi Arabia is striving to become the third largest AI market in the world, while leading players in France and Israel are pledging to lead in specialized AI applications. With its immense population and growing emphasis on education, India is shaping to be among the primary suppliers of engineering and computer science talent.
The international order is becoming more multi-polar, and the AI world is no exception. Instead of trying to "win the AI race" at any cost against its rival, both the U.S. and China should build bridges and seek common ground with friends and rivals alike.
This essay is adapted from the authors’ forthcoming book, Geopolitics of Artificial Intelligence, to be published in 2026 by Cambridge University Press as part of its Elements series.
