The U.S. and U.N. Part Ways on AI Regulation
The United States and the United Nations are diverging in their approach to AI governance. Recently, at the U.N. General Assembly, the U.S. government explicitly rejected proposals to establish a global framework for artificial intelligence governance, highlighting a major disagreement with the international community on AI regulation. Michael Kratsios, the U.S. AI policy chief, emphasized at a U.N. Security Council meeting that the U.S. “completely rejects” any attempts by international organizations to exert “centralized control and global governance” over AI, insisting that the future of AI “lies not in bureaucratic management but in national independence and sovereignty.” Meanwhile, among the 193 U.N. member states—including China—the majority support establishing a framework for international cooperation. This reflects a growing global divide in technology governance, entering a new stage of fragmentation.
Rising Intensity of AI Competition
In recent years, China has rapidly emerged in the AI field, putting pressure on the traditional technological powerhouse, the U.S. China is leveraging its massive data resources to accelerate domestic innovation and promote algorithms abroad. Tech giants like Baidu, Alibaba, Tencent, and Huawei are driving cutting-edge innovations, developing advanced facial recognition systems, language-processing tools, and other technologies. At the recently held World Artificial Intelligence Conference in Shanghai, Chinese Premier Li Qiang proposed establishing a global AI cooperation organization to promote multilateral, open-source collaboration—signaling Beijing’s ambition to expand China’s influence in global geopolitics.
Recent data shows China leads the U.S. in AI patent applications by nearly tenfold, and China’s AI research output has surpassed the combined total of the U.S., the 27 EU countries, and the U.K. Facing this reality, the U.S. insists on maintaining its technological advantage through sovereign control, while China calls for strengthened international cooperation. The European Union, meanwhile, is pursuing a “third way” via its AI Act, formally released in July 2024, hailed as “the world’s first comprehensive AI law.” This multipolar governance model reflects a global AI landscape trending toward regionalization and fragmentation rather than unified international standards. Establishing an international AI organization is “one of the most critical issues of our time,” but achieving this goal requires direct negotiation and cooperation between the U.S. and China—a prospect that currently looks bleak.
In the West, concerns are growing that China’s dominance could shape global technology standards and governance, potentially exporting its ideology and weakening the influence of democratic nations in global tech governance. China has actively increased its participation in international standard-setting, particularly in developing countries, promoting AI systems like facial recognition with low cost and high efficiency. The most representative case is TikTok, which faced scrutiny over national security and data privacy while expanding abroad. The Trump administration restricted its use on government devices and demanded its sale to U.S. companies. With 170 million users in the U.S., over half the population, TikTok’s expansion prompted the White House to launch an “Action Plan” to enhance domestic technology and counter China’s influence. “Just as we won the space race, the U.S. and its allies must win this AI race,” the White House stated in the Action Plan.
Diverging Trends in 2025
This divide has become even clearer in 2025. According to Stanford HAI’s 2025 AI Index report, U.S. private AI investment reached $109.1 billion—almost 12 times China’s $9.3 billion—highlighting an innovation model dominated by Western capital markets. The U.S. produced 40 top-tier AI models, leading globally. However, China is rapidly closing the performance gap; a RAND report predicts that Chinese AI models will match U.S. capabilities by 2025. China focuses on “AI+” vertical applications, such as agricultural AI advisors and medical diagnostic systems.
The divergence stems from systemic differences: the West relies on market competition and open innovation, rejecting the U.N.’s global governance framework and emphasizing sovereign control to preserve its technological advantage. China, on the other hand, supports open-source models like DeepSeek through national funds (around ¥60 billion) and local government initiatives, emphasizing low-cost, scalable deployment. Discussions on X suggest that China’s “embodied AI” (robots) could dominate global value creation by 2030. The U.S. pursuit of AGI is disruptive, likened to an atomic bomb, whereas China’s application-driven approach is pragmatic—addressing export bans and achieving 70% domestic chip production. The result is a “dual-track” global AI ecosystem: Western high-end innovation and Chinese industrial empowerment, with China trailing only 6–12 months behind in model development.
Different Visions, Different Paths
Although competition between the U.S. and China in AI is intensifying, their development paths are increasingly distinct. The U.S. is investing hundreds of billions of dollars, consuming thousands of megawatts of energy, and racing to surpass China in the next AI evolutionary leap. Some view this leap as powerful enough to rival an atomic bomb in its impact on the global order. Since the launch of OpenAI’s ChatGPT nearly three years ago, Silicon Valley has poured vast resources into pursuing the “holy grail” of AI—Artificial General Intelligence (AGI) capable of rivaling or exceeding human thought.
China, by contrast, is running a different race. Amid growing concerns about an AI bubble, China has said little about AGI and is instead pushing its tech industry to “focus firmly on applied fields”—developing practical, low-cost tools that boost productivity and are easy to commercialize, countering Silicon Valley’s pursuit of superintelligent AI.
Currently, U.S. tech companies are developing pragmatic AI applications. For instance, Google connects its Pixel smartphones to the internet for real-time translation; U.S. consulting firms use AI agents to create presentations and summarize interviews; other companies improve drug development and food delivery. Unlike the largely laissez-faire approach in the U.S., China is actively supporting its vision. In January, China established a National AI Fund totaling ¥60.06 billion, focusing on startups, followed by local government and state-owned bank initiatives, along with city-level AI development plans under the “AI+” program.
While Chinese companies are releasing their best models openly, U.S. companies prefer to keep “shiny new products” proprietary. Meta, Google, and OpenAI compete heavily to secure talent, data centers, and energy. The U.S.-China Economic and Security Review Commission (USCC) even recommended a “Manhattan Project”-style initiative to fund AGI development and ensure U.S. leadership. However, given uncertain returns on large-scale investment, the U.S. path may not be wiser. Ultimately, like the internet’s bubble and years of development, AI competition could take decades to determine winners.
This divergence affects not only technology but also the global economy, military balance, and societal change. AI is expected to contribute $15.7 trillion to global GDP by 2030. Western innovation drives high-value industries like cloud services and pharmaceuticals, but 95% of companies see no ROI, raising bubble concerns. China’s application model accelerates manufacturing transformation, reduces software costs, and affects U.S. software market valuation by $1 trillion. Supply chain fragmentation increases global costs by 2–3%, forcing developing countries to choose sides: U.S. security vs. China’s affordability.
The IMF warns that AI may widen the wealth gap, affecting 40% of jobs globally, benefiting Western white-collar workers while low- and medium-skill labor faces unemployment risks. The path leads to a multipolar economy, with China exporting AI systems to developing countries, weakening Western influence.
In military terms, AI divergence changes the global landscape. The U.S., through the AUKUS alliance, maintains air superiority and nuclear stability, but China’s hypersonic missiles and AI drone swarms threaten the Taiwan Strait. RAND simulations suggest U.S. missile stockpiles could be depleted in 72 hours, while Chinese AI electronic warfare disrupts radar systems. Chinese military AI investments escalate U.S.-China competition, and the U.S. Action Plan treats AI as a space-race-like challenge.
Global conflict is transforming. AI lowers attack costs, as seen in cyber and drone warfare in Ukraine. NATO faces Russia-China alliances; AI weaponization heightens South China Sea tensions, yet interdependence prevents full-scale war. GIS reports predict AI will reshape geopolitics by 2030, and the West must win the AI race to maintain advantage.
Societal change is also impacted. The West emphasizes transparency and ethics (e.g., EU AI Act), while China strengthens surveillance and efficiency. AI replaces routine work, with Western high-wage jobs benefiting from augmentative AI; China’s application model accelerates social control (e.g., police AI dispatch). Ethical divergence arises as the West worries about China exporting ideology, while China promotes digital collectivism. Paths lead to social polarization, a widening digital divide in developing countries, and fragmented AI ethics standards. Discussions on X suggest the AI “cold war” is forming new blocs, requiring policy buffers for employment transitions.
The Oligopoly Era Arrives
The AI industry’s competitive landscape is fundamentally changing. In the AI chatbot market, ChatGPT still holds a 60.6% share, Google Gemini 13.4%, Microsoft Copilot 14.1%, and other competitors under 7%. This concentration allows resource-rich tech giants to continually expand their lead. The global AI market is expected to grow from $391.7 billion in 2025 to $1.81 trillion by 2030, with a 35.9% CAGR—surpassing the cloud computing boom of the 2010s. Microsoft, IBM, AWS, Google, and NVIDIA collectively hold 42–48% of the market.
Notably, the AI industry is seeing a divergence in technological approaches. Google’s vertically integrated AI ecosystem—from TPU chips to application services—challenges NVIDIA’s dominance in AI chips. Microsoft recently announced it would use both Anthropic and OpenAI technologies in Office 365, ending its exclusive reliance on OpenAI—a “don’t put all eggs in one basket” strategy that reduces technical risk and improves user experience. This multi-vendor approach is emerging as a trend among major tech firms.
In response, OpenAI seeks independence, planning to mass-produce its own AI chips with Broadcom by 2026, reducing reliance on Microsoft Azure. OpenAI also launched a job platform challenging LinkedIn. Anthropic, through its Microsoft partnership, gains access to 430 million Office 365 users. Its Claude Sonnet 4 model already surpasses GPT-5 in some tasks, providing a differentiated advantage in enterprise markets.
In 2025, AI investment reached $364 billion, dominated by U.S. giants, though China is catching up. U.S. moves: Microsoft invests $80 billion in AI infrastructure and ends exclusive OpenAI reliance, adopting multi-vendor strategies (e.g., Anthropic); Google TPU integration challenges NVIDIA, raising market value 800%; OpenAI pursues independence and self-produced chips by 2026. Export controls maintain U.S. advantage, but Chinese open-source alternatives undercut profits.
China plans $98 billion in AI investments, including Huawei’s Ascend chip mass production, Baidu and Alibaba AI cloud deployment, and DeepSeek open-source models potentially disrupting Western profits. These efforts aim to circumvent U.S. localization bans, export low-cost hardware, and enhance domestic party-controlled applications.
Looking back, history repeats itself. The current AI competition resembles the cloud computing battles of the past—an oligopoly is forming. Few giants, with strong capital and computing power, will define the market, and the “winner-takes-all” principle will likely reappear. True winners often emerge over two to three generations; for example, Google is the third generation of search, Facebook the third of social networks. Who will ultimately dominate in brand building, independence, and market share remains uncertain.
Future Outlook
The diverging paths of the AI race suggest long-term uncertainty: Western innovation vs. Chinese applications. Globally, risks must be balanced. In 2025, a bubble may burst, but as with the internet, winners may only emerge after decades. Cooperation may be key; otherwise, fragmentation could exacerbate geopolitical tensions.