The US-China Race for AI Supremacy
While the trade war between the US and China may have stalled, the competition for technological dominance is intensifying. Both nations are vying for leadership in artificial intelligence (AI), with significant geopolitical and productivity benefits at stake. A pressing question looms: will China surpass or match the US in AI capabilities?
Trump’s Policies: A Catalyst for China’s AI Advancement
The driving force behind this trend is a series of policies implemented by the administration of US President Donald Trump. Trump’s presidency marks a stark departure from the openness that has underpinned US technological leadership for decades. Measures aimed at bringing innovation back to the US could instead pave the way for Chinese dominance.
Historical Context: The Evolution of Digital Economies
In the 1990s, the US led the internet revolution, dominating the crucial “zero to one” phase by swiftly transferring lab innovations to the market. This fueled what many hailed as the “new economy,” marked by rapid growth, strong productivity gains, and low inflation. China, initially a follower, later injected significant dynamism into the digital economy by expanding its innovative technologies.
China’s digital development occurred in three phases:
- Copying and Following (Mid-1990s to early 2000s): Chinese companies imitated US models, launching web portals and online services that fueled explosive user growth.
- Localization and Improvement (2005-2015): As China’s digital ecosystem matured, local tech companies leveraged deep user knowledge and market conditions to refine their services. Platforms like WeChat and Taobao not only adapted US concepts but surpassed Western counterparts such as WhatsApp and eBay in the Chinese market.
- Innovation with High Impact (Last 10 years): Chinese tech companies have transitioned from imitation to innovation, pioneering new digital models and outpacing foreign competition. TikTok by ByteDance exemplifies this, placing China at the forefront of online culture, reshaping social networks, and compelling US firms like Meta to catch up.
This pattern is already evident in fields like renewable energy and electric vehicles (EVs), and AI will follow suit. After ChatGPT’s launch in late 2022, which likely marked AI’s mass adoption era, China quickly demonstrated its ability to replicate Western models.
DeepSeek’s January release signaled China’s entry into the localization and improvement phase, as its R1 model was 30 to 50 times cheaper than OpenAI’s. By February, the performance gap between top Chinese and US models narrowed to 1.7%, compared to 9.3% in 2024. While ChatGPT took two months to reach 100 million active users, DeepSeek achieved this milestone in just seven days.
China’s Advantages
One of China’s key advantages is its abundant pool of engineers. The country graduates four times more STEM students than the US, reflecting a strong work ethic and pragmatic mindset suited for complex, practical optimization—as seen in DeepSeek’s architecture.
With over 1.0 billion internet users and a diverse industrial base, China offers unparalleled conditions for deploying, testing, and refining AI applications. China accounts for nearly 30% of global manufacturing output, generating massive amounts of data. In 2019, its manufacturing sector produced 1.812 petabytes of data, estimated to reach 2.435 PB by 2024.
Energy is another critical factor. In 2023, China generated approximately 9.456 terawatt-hours of electricity—32% of the global total and more than double the US production of 4.178 TWh—giving it a significant edge in powering large-scale AI data centers essential for widespread AI adoption.
US Position Weakened by Trump’s Cuts and Restrictions
Trump’s policies further undermine the US position in the AI race through funding cuts for research and immigration restrictions. For instance, in February 2021, the Trump administration dismissed 170 employees, including AI experts, from the National Science Foundation and proposed a more than 50% budget cut for the agency.
These cuts, along with delays in NIH funding allocations and freezing of approximately $2.2 billion in federal grants to Harvard University, risk stalling foundational research and hindering AI innovation. Meanwhile, restrictive immigration policies may make it difficult for the US to attract and retain global talent, potentially triggering a reverse brain drain as qualified Chinese tech workers return home for lucrative positions in a booming sector.
Although the Trump administration has supported large-scale infrastructure initiatives like OpenAI’s proposed $500 billion AI data center, Stargate, such projects risk reinforcing the dominance of big tech and stifling necessary innovation for transformative technological advancements.
The core issue lies in the US distancing from economic openness. As US companies like OpenAI become increasingly closed, Chinese firms adopt open-source strategies. Meanwhile, Trump’s trade and immigration policies deter global talent and international collaborators, while China actively markets its low-cost AI models among trading partners.
Despite internal challenges exacerbated by US trade restrictions limiting access to advanced semiconductors, Chinese policymakers must strike a delicate balance between fostering innovation and implementing strict data controls. Although neither side has an easy path to AI dominance, the MAGA agenda inadvertently may help make China “great again.”
About the Authors
Qiyuan Xu, Principal Researcher and Deputy Director at the Institute of World Economics and Global Policy, Chinese Academy of Social Sciences, is the author of several books, including “Reshaping the Global Industrial Chain: China’s Choices.”
Wang Yaqiang is a professor at the Lee Kuan Yew School of Public Policy, National University of Singapore.
Copyright: Project Syndicate, 1995 – 2025
www.project-syndicate.org
