U.S. Struggles to Counter China's AI Advances Amid Trade-Offs

The United States faces a strategic dilemma in its efforts to maintain artificial intelligence leadership over China, as existing export controls prove insufficient and new restrictions risk economic retaliation, according to a Bloomberg News report.

Despite years of semiconductor export restrictions targeting advanced AI chips, Chinese companies have steadily closed the capability gap with American frontier models. Firms including DeepSeek and Moonshot have developed AI systems that rival offerings from Anthropic and OpenAI, often at significantly lower costs that threaten Silicon Valley's market position.

Why it matters

The AI competition between Washington and Beijing increasingly defines the broader technology rivalry between the world's two largest economies. Yet the U.S. government's ability to slow China's progress is constrained by economic interdependencies—particularly Beijing's control over rare earth minerals essential for advanced weapons systems and China's role in data center supply chains. These constraints force American policymakers to balance national security objectives against immediate economic vulnerabilities, limiting the effectiveness of unilateral technology restrictions.

The Distillation Problem

U.S. officials and AI developers have accused Chinese companies of using "distillation"—a technique where outputs from advanced models train less sophisticated systems—to improperly replicate American AI capabilities. Anthropic recently claimed that Moonshot AI routed nearly 300,000 user requests through its Claude system and presented the results as original work.

Closing this loophole proves difficult in practice. Chinese users can remotely access American semiconductors through cloud data centers in Southeast Asia, a workaround that expanded after Trump administration trade policies alienated regional partners. Proposed legislation to restrict such cloud access would require monitoring data centers across multiple countries, presenting significant enforcement challenges.

Rare Earth Leverage Limits Options

Beijing's dominance in rare earth production gives China retaliatory leverage that constrains U.S. policy options. Trump and Xi Jinping negotiated a truce guaranteeing rare earth flows in exchange for reduced tariffs—an arrangement likely to be extended during Xi's upcoming White House visit on September 24.

"Washington has to tread carefully now that Beijing has sharper retaliatory tools, such as export controls on rare earths," said Kyle Chan, a fellow at the Brookings Institution's John L. Thornton China Center. "Also, the American AI boom relies on many data center components from China."

The Trump administration has maintained export controls on the most advanced chips while showing openness to selling China less restricted semiconductors like Nvidia's H200s. Officials have avoided more aggressive measures such as entity-list designations for companies engaged in systematic distillation.

Global Competition Expands

The AI rivalry extends beyond bilateral tensions. Huawei Technologies has bid to build AI data centers for Egypt's government, prompting a competing offer from Washington. China increasingly dominates manufacturing supply chains for electric vehicles, clean energy equipment, and drones—sectors where the U.S. has imposed restrictions on national security grounds.

Daniel Remler, a senior fellow at the Center for a New American Security who led AI policy at the State Department under the Biden administration, noted that "leaders and senior policymakers on both sides have to see AI in the context of every other issue they have to contend with, from the war in Iran to rare earths."

AI discussions are expected to feature prominently in the September 24 summit between Xi and Trump, with separate technical talks planned around the meeting. China has dismissed U.S. theft allegations as groundless and demanded removal of "containment measures" as a precondition for meaningful cooperation.

These details were first reported by Bloomberg News.