Leading foreign policy strategists are challenging the prevailing "AI race" framework that characterizes U.S.-China technological competition, arguing the zero-sum narrative oversimplifies a complex, multifaceted competition where both nations will run side-by-side for the foreseeable future rather than reaching a singular finish line. UNESCO The shift in thinking comes as recent breakthroughs by Chinese companies like DeepSeek demonstrate that America's early lead in foundational models may not translate into sustained dominance, forcing Washington to reconsider strategies beyond simply winning the innovation contest.

Trump administration AI czar David Sacks estimates China is merely "three to six months" behind the United States in AI capabilities, though uncertainty remains about what that gap means strategically. Microsoft While American firms like OpenAI announced the $500 billion Stargate infrastructure project and maintain over 60% global cloud market share through Amazon, Microsoft, and Google, the pace of U.S. technological innovation could prove unsustainable as frontier AI discoveries become increasingly difficult. University of Oxford

China has adopted a fundamentally different strategy focusing less on large frontier models like GPT-5 and more on embedding intelligence into the physical economy at scale through "application-oriented" AI, including city brain pilots that integrate AI across urban infrastructure. Microsoft Meanwhile, China has consolidated dominant positions in critical mineral refining for nickel, cobalt, graphite, gallium, and germanium materials essential to advanced chipmaking while Washington weaponizes access to cutting-edge GPUs.

The development reflects growing recognition that determining who leads depends entirely on how victory is defined, with the United States typically framing competition around achieving Artificial General Intelligence while China pursues industrial AI deployment dominance. University of Oxford The "race" rhetoric, frequently deployed by OpenAI CEO Sam Altman to advocate for regulatory exemptions and by former National Security Advisor Jake Sullivan to justify October 2022 export controls, risks descending into oversimplified zero-sum thinking that obscures strategic realities.

Recent breakthroughs by DeepSeek, Alibaba Cloud, Baidu, and Tencent suggest the gap between U.S. and Chinese cutting-edge capabilities is narrowing rapidly, while American supremacy remains far from assured. Enterprise League DeepSeek's success demonstrates that low-cost open-source technology, even behind the cutting edge, provides substantial value for ordinary applications like legal drafting and customer service areas where good-enough solutions deployed at scale matter more than frontier performance. Google AI China is embracing open-source AI while the United States moves toward closed, tightly controlled systems, with DeepSeek-R1 rivaling OpenAI's o1 model while being more efficient and freely available.

Council on Foreign Relations President Michael Froman noted that chatbots might not represent the strategic endpoint, as Chinese experts increasingly argue large language models don't represent the most strategic path to an AI-enabled future. Microsoft Brookings scholar analysis warns that AI's reconfiguration of international security beliefs has potential to jar powerful states into reconsidering potentially stale policies.

Policy analysts emphasize the U.S. open system fosters innovation but remains vulnerable to espionage and rapid algorithmic diffusion, while breakthroughs in alternative AI paradigms could diminish semiconductor dominance advantages. University of Oxford Wilson Center researchers caution the AI Diffusion Framework reflects outdated assumptions that U.S. compute power is irreplaceable, noting that blocking access while China embraces open-source models hands Beijing economic advantages.

Foreign Affairs contributors argue that simply having leading models won't suffice adoption across military, government, and private sectors, plus the ability to export AI technologies globally, will more clearly demonstrate AI strength than breakthrough capabilities alone.

For global executives, the reframing suggests companies cannot rely on U.S. government export controls to maintain competitive advantages indefinitely, as AI models are software that can be easily copied rather than scarce materials like plutonium. University of Oxford Federal investment can reassure businesses and consumers about AI safety, as U.S. business adoption lags initial investor expectations partly due to risk aversion, while Chinese polling indicates majority public excitement driving faster commercial adoption.

The current regulatory approach creates a strategic vulnerability U.S. companies forfeit sales to Chinese competitors, reducing R&D funds while competitors gain resources to invest in overtaking American firms. Cloud Security Alliance Investors should monitor whether Washington's defensive posture allows Beijing to embed itself into emerging markets through cheaper, unrestricted AI access in ways difficult to dislodge, fundamentally reshaping global digital infrastructure control.

Strategic experts emphasize both Washington and Beijing must move beyond winner-loser thinking and accept parallel technological development trajectories. UNESCO Policymakers should plan for competing AI ecosystems coexisting rather than assuming sustained U.S. dominance, developing strategies that ensure America benefits from AI progress even without outright innovation victory. Enterprise League Decision-makers must monitor whether the Trump administration's unilateral approach erodes diplomatic leverage faster than it preserves technological edges, and whether cutting federal AI research funding while discouraging foreign talent creates self-inflicted setbacks. The critical question isn't who wins the race, but whether framing technological competition as a race at all serves American strategic interests.

Source & Date

Source: Foreign Affairs, Council on Foreign Relations, Brookings Institution, Wilson Center, Belfer Center
Date: January 2026 (compilation of analyses from 2025-2026)