The evolving landscape of artificial intelligence is dramatically influencing global power dynamics. Over the coming weeks, experts from leading technology and financial publications will engage in a series of discussions exploring various facets of the generative AI revolution and its geopolitical implications.
In this dialogue, John Thornhill, a technology columnist and Innovation Editor, teams up with Caiwei Chen from MIT Technology Review to analyze the intensifying technological rivalry between Silicon Valley and Beijing.
Assessing the AI Power Struggle: China’s Rising Influence
John Thornhill’s Perspective:
From an international vantage point, China’s ascent as a dominant AI force in the 21st century appears inevitable.
Western observers often emphasize America’s commanding lead in semiconductor technology, pioneering AI research, and massive investments in data infrastructure. Legendary investor Warren Buffett famously advised, “Never bet against America,” underscoring the United States’ unparalleled history as a cradle for innovation and human potential over the past two centuries.
Yet, China now possesses the resources, strategic intent, and societal mobilization capacity to challenge this dominance profoundly. Its ability to harness nationwide efforts to accelerate AI development and deployment suggests that underestimating China could be a critical miscalculation.
Empirical data underscores this shift. According to Stanford University’s 2025 Artificial Intelligence Index Report, China led global AI research citations in 2023 with 22.6%, surpassing Europe’s 20.9% and the US’s 13%. Moreover, China accounted for nearly 70% of AI patent filings worldwide that year. Although the US still holds a majority in the top 100 most cited AI publications (50 compared to China’s 34), its share is steadily diminishing.
Talent distribution also reflects this trend. The US Council of Economic Advisers reported that in 2019, 59% of the world’s leading AI researchers were based in the US, while China accounted for 11%. By 2022, these figures shifted to 42% and 28%, respectively, indicating a narrowing gap.
Policy changes, such as the Trump administration’s stricter H-1B visa regulations, may accelerate the repatriation of Chinese AI experts, further bolstering China’s talent pool.
Technologically, US institutions produced 40 of the most influential AI models in 2024, compared to China’s 15. However, Chinese researchers have demonstrated remarkable efficiency, with models like the open-source DeepSeek-V3 and Alibaba’s Qwen 2.5-Max outperforming US counterparts in algorithmic optimization.
China’s Edge in AI Application and Hardware Innovation
China’s true strength may lie in the practical application of AI technologies. Recent data from Air Street Capital reveals that China has surpassed the US in monthly AI model downloads, reflecting widespread adoption. In sectors such as fintech, e-commerce, and logistics, China already leads in AI integration.
Looking ahead, hardware innovations-particularly in drones and industrial robotics-are poised to be game-changers. As AI research increasingly focuses on embodied intelligence, China’s manufacturing prowess offers a significant advantage.
Tech analyst Dan Wang, author of Breakneck, highlights China’s engineering-driven approach to manufacturing innovation. While this has propelled technological and economic growth, it also coincides with intensifying social repression, which could complicate the country’s broader ambitions.
John Thornhill invites Caiwei Chen to weigh in on how China’s social governance might influence its AI trajectory.
China’s AI Ambitions: Balancing Innovation and Control
Caiwei Chen’s Response:
Hello John,
You correctly note that the US maintains a lead in cutting-edge AI research and infrastructure. However, defining “victory” in AI is complex. Jeffrey Ding’s analysis in Technology and the Rise of Great Powers suggests that the widespread societal integration of AI technologies often determines long-term dominance. In this respect, China is well-positioned to excel-though perhaps “technological murder” is an overstatement.
China’s greatest challenge remains its semiconductor industry. Export controls have severely limited access to advanced GPUs, forcing reliance on secondary markets and refurbishment of existing hardware. Despite expanding domestic chip initiatives, a performance gap persists at the highest levels.
These constraints have spurred Chinese firms to innovate differently: emphasizing shared computing resources, maximizing efficiency, and promoting open-weight AI models. For instance, DeepSeek-V3’s training consumed only 2.6 million GPU-hours, significantly less than comparable US models. Alibaba’s Qwen series ranks among the most downloaded open-source models globally, while companies like Zhipu and MiniMax develop competitive multimodal and video AI systems.
China’s industrial policies facilitate rapid transition from research to real-world deployment. Local governments and major corporations are already integrating AI-driven reasoning models into public administration, supply chains, and financial services.
Education also plays a pivotal role. Leading Chinese universities have embedded AI literacy into their curricula, preparing students ahead of labor market demands. The Ministry of Education plans to introduce AI training across all school levels, fostering early familiarity. While the term “engineering state” may not fully capture China’s nuanced relationship with technology, decades of infrastructure development and centralized coordination have enabled swift, large-scale adoption with comparatively minimal social resistance. This widespread use accelerates iterative improvements.
Interestingly, the 2025 Stanford HAI AI Index found Chinese respondents to be the most optimistic globally about AI’s future, outpacing sentiments in the US and UK. This optimism persists despite China’s recent economic slowdown-the first in over twenty years-highlighting AI’s perceived role as a catalyst for renewed growth. Whether this enthusiasm endures amid economic challenges remains uncertain.
Regarding social control, a new generation of Chinese AI entrepreneurs exhibits a global outlook, seamlessly navigating between Silicon Valley innovation hubs and international investment arenas like Dubai. Fluent in English and versed in global venture capital dynamics, these founders are building companies with inherently transnational ambitions, moving beyond the constraints of a solely domestic identity.
While the US continues to lead in rapid experimentation and innovation speed, China’s influence on the everyday integration of AI-both domestically and internationally-could be profound. Speed alone does not equate to ultimate dominance.
Reflections on the US-China AI Rivalry
John Thornhill’s Rejoinder:
Caiwei, you make an excellent point distinguishing speed from supremacy, and I agree that “technological murder” was an overstatement. Your emphasis on China’s success with open-weight AI models versus the US’s preference for proprietary systems highlights a fundamental divergence-not just between two nations, but between contrasting technological deployment philosophies.
Even OpenAI’s CEO, Sam Altman, acknowledged earlier this year that the company had “been on the wrong side of history” regarding open-source strategies, signaling a potential shift in approach. Tracking how this dynamic unfolds will be fascinating.
Additional Insights on the US-China AI Competition
- Explorations into how generative AI is becoming part of everyday life reveal diverse adoption patterns across regions.
- Analyses question how long Nvidia can maintain its GPU dominance amid rising Chinese competitors.
- Emerging AI applications in consumer devices, such as interactive toys and companion robots, are gaining traction in both China and the US.
- China’s once-rapid data center expansion has encountered obstacles, prompting stakeholders to adapt strategies amid shifting sanctions and AI demands.
