In September 2026, a growing rift over artificial intelligence development has erupted between the world’s two largest economies, sparked by a leading American AI chief’s urgent call for safety slowdowns and new export restrictions that have drawn sharp pushback from Beijing. The controversy began when Dario Amodei, CEO of leading AI developer Anthropic, published a weekend essay sounding the alarm over the accelerating pace of frontier AI development, warning that unregulated progress could bring catastrophic global risks within a year.
Amodei warned that the most powerful next-generation AI systems could, in as little as six to 12 months, build a persistent botnet capable of seizing full control of the global internet, causing hundreds of billions of dollars in economic damage. Without binding safety guardrails, he argued, risks will only grow exponentially as models grow more powerful. He called for a deliberate slowdown in the development of the most advanced frontier AI systems, giving governments, researchers and regulators time to implement robust risk management frameworks.
His stance quickly won backing from other major U.S. tech figures, including OpenAI CEO Sam Altman and Tesla and SpaceX CEO Elon Musk. Beyond safety concerns, Amodei framed Chinese AI advancement as a major strategic threat to the U.S. and the world, urging the White House to maintain strict restrictions on exports of cutting-edge AI chips and chip manufacturing equipment to China. He also called for new measures to crack down on what he claims is unauthorized “distillation” of U.S.-developed AI models by Chinese research labs.
U.S. President Donald Trump, however, has pushed back on calls to slow AI development, dismissing many safety warnings as overblown while doubling down on his framing of AI as a zero-sum global competition. “We are ahead of China in AI. We are the most advanced country in the world, and frankly, I want to keep it that way — because whoever wins the AI race wins everything,” Trump stated Sunday, aligning with his administration’s long-standing hardline approach to technological competition with Beijing.
China has hit back against Amodei’s proposals, accusing the U.S. of using AI safety as a cynical pretext to lock in its long-held technological dominance. On Monday, Chinese Foreign Ministry spokesperson Guo Jiakun called for a global approach to AI governance that is open, inclusive and collaborative. “Stoking unfounded fears, pushing confrontation and engaging in malicious competition will only disrupt global AI governance processes, and serve no one’s interests,” Guo said. The Chinese state-backed *Global Times* went further, labeling Amodei’s plan a Cold War-era tactic designed to stifle Chinese AI progress and preserve U.S. technological hegemony, calling the emerging “quiet AI Cold War” hypocritical and short-sighted.
The current clash is unfolding against a well-documented backdrop of intensifying AI competition between the two powers. Experts note that while the U.S. still maintains leading positions in key segments of the AI race, China is rapidly closing the gap. Data from Stanford University’s 2026 AI Index Report shows U.S. private companies invested $285.9 billion in AI development in 2025, dwarfing China’s $12.4 billion in private investment. The U.S. also retains a clear lead in the production of the advanced semiconductors required to train cutting-edge large language models.
Despite sweeping U.S. export restrictions on top-tier AI chips, Chinese firms have already proven they can develop competitive large AI models at a fraction of the cost of their U.S. rivals. In early 2026, Chinese AI startup DeepSeek made global headlines when it revealed it had trained its latest flagship model using less than $6 million worth of computing power from Nvidia H800 chips — a tiny fraction of the $100 million OpenAI reportedly spent to train its GPT-4 model back in 2023. The announcement sent shares of major U.S. tech firms tumbling, as it upended widespread assumptions about the cost barriers to cutting-edge AI development.
The U.S. has enforced broad restrictions on Chinese access to advanced semiconductors and chip manufacturing technology for years, justifying the limits on national security grounds, arguing advanced AI chips could boost China’s military, intelligence and surveillance capabilities. The Trump administration has maintained bans on the most cutting-edge chips, including Nvidia’s Blackwell processors, while allowing limited exports of lower-power chips to the Chinese market.
China, meanwhile, holds key advantages across other segments of the AI supply chain. It dominates global production of the critical rare earth metals required for advanced semiconductor manufacturing, and it generates more than twice as much electricity as the U.S., giving Chinese AI data centers access to abundant low-cost energy to power energy-intensive AI training and operations. In late 2025, China tightened export controls on five critical rare earth metals and restricted exports of specialized refining equipment for these resources, a move widely seen as a countermeasure to U.S. chip restrictions.
Beyond the commercial and technological competition, both powers have faced scrutiny for their growing use of AI in military and surveillance applications, despite both framing their own AI policies as responsible. In June 2026, the Trump administration ordered a rapid expansion of AI use across U.S. intelligence and military operations. In March, the U.S. military confirmed it had deployed advanced AI tools to process intelligence during operations against Iran, though it emphasized human officials retain final authority over target selection. That confirmation came amid calls for an independent investigation into a bombing of a school in southern Iran that killed 156 people, most of them children.
Multiple independent reports have also documented extensive Israeli reliance on AI for target selection, surveillance and operational planning during its ongoing military campaign in Gaza, which has killed more than 73,800 Palestinians since October 2023 and left most of the territory in ruins. A 2025 report from United Nations Special Rapporteur Francesca Albanese named major U.S. AI and tech firm Palantir as one of the companies supporting Israel’s military campaign and displacement of Palestinian populations in violation of international law. Google and Amazon have also faced widespread backlash over their $1.2 billion Project Nimbus deal, struck in 2021, to supply cloud and AI-powered surveillance infrastructure to the Israeli government and military.
Notably, Anthropic itself has clashed with the U.S. Pentagon over military AI use, refusing to allow its models to be used for fully autonomous weapons or large-scale domestic surveillance.
China has publicly criticized the growing militarization of AI, warning against ceding life-or-death decision-making power to algorithms. But independent investigations have also raised red flags over Chinese military and surveillance use of AI: a July 2026 Reuters investigation found researchers with ties to the Chinese military have used outputs from leading U.S. AI models to enhance Chinese defense capabilities, including for surveillance applications. A 2025 Associated Press investigation documented how AI has become a core component of a massive Chinese mass surveillance system targeting dissidents and ethnic minority groups including Uyghurs and Tibetans, and found U.S. tech firms played a key role in supplying core technology for that system. In February 2026, both the U.S. and China refused to sign a joint declaration regulating military AI use at an international military AI summit in Spain.
AI industry analysts say the high-profile calls for a slowdown from top tech leaders have raised new questions about whether the private sector can or will regulate the rapidly advancing technology responsibly. Aya Ibrahim, senior visiting fellow at the AI Now Institute, told Al Jazeera that warnings from tech leaders come as private firms face intense financial pressure to roll out ever-more powerful AI systems as quickly as possible to capture market share. “So much money is riding on getting these new systems to market and performing well, and there is growing doubt that companies can both deliver profit and manage risk,” Ibrahim explained.
She also questioned the dominant framing of AI development as a geopolitical and commercial race, arguing this narrative pushes firms to prioritize speed over safety and accountability. “The race narrative is inherently problematic, because it pushes us toward a race to the bottom,” Ibrahim said. “You might win, but what do you actually win, and at what cost?”
