Two U.S. officials familiar with the preparations say AI will be “at the forefront” of talks, though they caution that concrete agreements are unlikely. The expectation, instead, is a cautious exchange shaped by rising competition and mutual suspicion.
The timing reflects how sharply the global AI race has escalated. One senior analyst described the current moment as increasingly resembling “a Cold War-style nuclear arms race,” driven by breakthroughs in frontier models and their expanding security implications.
Frontier Models, Security Fears, and a Widening Tech Divide
Tensions have intensified following the release of advanced systems such as Anthropic’s Mythos model by Anthropic. Analysts say China’s exclusion from early access to the system has deepened concerns in Beijing that cutting-edge AI could be used in ways that expose vulnerabilities in domestic software and financial infrastructure.
One researcher noted that Mythos had identified “thousands” of major vulnerabilities across operating systems and other digital infrastructure, prompting governments and financial institutions worldwide to reassess their cybersecurity posture.
The scale of the concern has widened the debate beyond innovation and into national security. White House officials acknowledged that systems like Mythos have made it necessary to maintain a “channel of communication” with Beijing to reduce the risk of miscalculation or unintended escalation.
At the same time, firms such as IDC China warn that excluding Chinese companies from frontier models risks creating what they call a “generational gap” in AI defense capabilities between China and the West.
Tech Heavyweights Join the Delegation
The presence of major industry figures suggests the conversation may extend beyond diplomacy into technical and commercial territory. Jensen Huang, chief executive of Nvidia, is expected to join the U.S. delegation alongside White House technology adviser Michael Kratsios.
Their participation has fueled speculation that advanced semiconductor systems—particularly Nvidia’s H200 chips—could feature in behind-the-scenes discussions, even as formal negotiations remain uncertain.
On the Chinese side, there are signals of interest in establishing structured engagement. Beijing has reportedly proposed a formal AI dialogue mechanism involving U.S. Treasury Secretary Scott Bessent and Chinese vice finance minister Liao Min. The proposal, first reported by the Wall Street Journal, reflects an effort to create institutional channels for AI-related coordination.
However, skepticism remains high in Washington, particularly given that neither participating agency has deep AI regulatory expertise and the Trump administration has only recently begun prioritizing AI safety oversight.
Calls for “Hotlines” and Crisis Mechanisms
Experts argue that as AI systems become more powerful, the absence of structured communication channels between the two superpowers increases systemic risk.
Kwan Yee Ng, head of international AI governance at Beijing-based Concordia AI, said a no-blame incident mechanism could be critical. “Getting senior Western figures to engage directly with China (on AI) has become increasingly difficult, though a positive signal from the Xi-Trump summit could change that,” Ng said.
Some analysts propose arrangements similar to past cybersecurity agreements or nuclear-era safeguards. One suggestion includes commitments to reduce AI-enabled malicious activity, while another envisions guardrails for frontier model deployment.
A military hotline already exists between the two countries, but U.S. officials have complained it is often unreliable, with calls going unanswered.
Sun Chenghao of Tsinghua University, who has participated in unofficial Track II discussions, said: “China likely hopes the U.S. will appropriately distinguish between AI governance and technological containment.”
Competing Visions of Control and Containment
Despite emerging dialogue efforts, fundamental disagreements remain over export controls and semiconductor access.
U.S. lawmakers continue to push stricter measures targeting China’s access to advanced chip supply chains, even as some restrictions have recently been eased under the Trump administration. The proposed MATCH Act has drawn strong objections from Beijing and could become a flashpoint during the summit.
Geopolitical strategist Reva Goujon of Rhodium Group said the moment is pivotal. “This is a really crucial window for Beijing to act and try to get the U.S. to commit to shutting it down,” she said.
Meanwhile, Chinese AI companies such as DeepSeek are increasingly emphasizing domestic chip development, though industry analysts note that U.S. export controls on semiconductor equipment continue to constrain China’s ability to scale production. Persistent computing shortages have forced some Chinese AI systems to limit user access.
Rising Accusations and Expanding Risk Scenarios
Beyond chips, tensions have also spilled into intellectual property disputes. Washington has accused China of large-scale theft of AI research and proprietary model technology from U.S. laboratories.
In response, Chinese state media has hardened its rhetoric. A recent article in a leading Communist Party publication described Western restrictions as moving beyond targeted measures into what it called a “systematic ecosystem blockade” against China.
Researchers warn that the stakes extend far beyond commercial competition. Advanced AI systems, they argue, could accelerate bioweapons research, destabilize financial markets, amplify cyberattacks and disinformation campaigns, and in extreme scenarios, behave unpredictably without human control.
As one expert summarized the core tension: “When one side sees AI as a proliferation risk to be contained and the other sees containment as an attack on a general-purpose technology, that makes it really difficult to find common ground.”