Trump and Xi Discuss AI Safety "Hotline" at Summit

President Trump welcomed Chinese President Xi Jinping to the White House Thursday for Xi's first U.S. state visit in more than a decade, with artificial intelligence sitting alongside trade, Taiwan, and the war in Iran as a central topic of discussion, according to Al Jazeera's live coverage of the summit.

Ahead of the talks, Treasury Secretary Scott Bessent confirmed the U.S. had proposed a new AI "notification mechanism" with China, effectively a hotline letting each country alert the other when an AI incident becomes serious enough to threaten national security, according to Al Jazeera's separate reporting on the summit's agenda. This connects directly to our earlier coverage of that same proposal surfacing during the UN Security Council session, where OpenAI, Anthropic, and Hugging Face executives briefed the council on AI risks just one day before this summit began.

This wasn't a routine diplomatic stop. It marked Xi's first White House visit in more than a decade, with Trump personally meeting him at the airport, a rare gesture underscoring the relationship's significance, according to Al Jazeera's reporting. The visit included a formal military ceremony with nearly 500 personnel and a state dinner attended by business and technology leaders. Beyond AI, the two leaders also discussed extending the U.S.-China trade truce, which Bessent said had been pushed back from its original November 10 expiration to January 10, giving both sides more time to negotiate a broader deal.

Al Jazeera's own data analysis, published the same day as the summit, offers a genuinely useful corrective to simple "who's ahead" framing. The U.S. dominates raw compute and spending: Goldman Sachs estimates U.S. hyperscalers, Amazon, Microsoft, Google, Meta, and Oracle, will spend roughly $764 billion on AI infrastructure in 2026, compared to about $102 billion from China's Alibaba, Tencent, Baidu, and ByteDance. The U.S. holds nearly three-quarters of the world's total AI computing power, versus just over 14% for China, according to Epoch AI. But China's spending is growing considerably faster, and the actual capability gap is narrower than the spending gap suggests, connecting directly to our earlier reporting on Kimi K3 rivaling Claude and ChatGPT.

A separate Moody's Ratings analysis adds real texture to this picture. While U.S. Big Tech outspends Chinese counterparts by roughly 5.6 to 1, the actual gap in installed data center computing capacity is only about 2 to 1, according to the Seoul Economic Daily's reporting on that analysis. Lower buildout costs, targeted state incentives, and cheaper green energy let Chinese firms secure considerably more computing power per dollar spent, with China's computing capacity reportedly growing roughly 19% annually, outpacing the U.S. That dynamic connects directly to our earlier reporting on China's secret weapon in the AI race, cheap, abundant energy letting Chinese firms build data centers considerably faster than their American counterparts.

This summit is worth understanding for any business evaluating exposure to U.S.-China trade and technology policy, since the proposed AI notification mechanism represents a genuinely concrete, if preliminary, bilateral safety measure emerging from the broader multilateral AI governance debate we've tracked closely throughout September, including King Charles III's own AI safety summit and the UN Security Council briefing from OpenAI, Anthropic, and Yoshua Bengio.

For businesses evaluating global AI infrastructure and vendor strategy, the genuine nuance in this compute-versus-spending data is worth internalizing directly: headline capital expenditure figures alone don't fully capture each country's actual AI capacity, and China's efficiency advantage in converting capital into usable compute is narrowing the practical gap faster than raw spending comparisons suggest.

What AI proposal did the US bring to the Trump-Xi summit?
Treasury Secretary Scott Bessent said the US proposed an AI "notification mechanism," effectively a hotline letting the US and China alert each other when an AI incident becomes serious enough to threaten national security.

Who is actually ahead in the US-China AI race?
The US holds a clear lead in raw computing power and spending, with nearly three-quarters of global AI compute versus China's roughly 14%, but China's AI spending and computing capacity are both growing faster, and analysts estimate leading Chinese models remain only months, not years, behind US frontier models.

Why does China get more computing power per dollar spent than the US?
Lower data center construction costs, targeted government incentives, and access to cheaper energy let Chinese firms build AI infrastructure more efficiently, meaning the actual computing capacity gap between the two countries is considerably narrower than the roughly 5.6-to-1 spending gap would suggest.

President Trump and Chinese President Xi Jinping discussed a proposed AI "notification mechanism," a bilateral hotline for national-security AI incidents, during Xi's first U.S. state visit in more than a decade, alongside trade, Taiwan, and the war in Iran. New data published the same day shows the US maintains a dominant lead in AI computing power and spending, roughly $764 billion versus China's $102 billion in 2026 hyperscaler spending, but China's capacity is growing faster and converting capital into usable compute more efficiently, narrowing the practical capability gap to roughly 2-to-1 despite a 5.6-to-1 spending disparity. The summit's AI discussions land directly amid an intense month of global AI safety diplomacy, following the UN Security Council briefing from OpenAI, Anthropic, and Yoshua Bengio just one day earlier.