When former U.S. President Donald Trump met China’s Xi Jinping in Beijing, the two leaders placed artificial‑intelligence supremacy at the centre of their dialogue. Trump pledged that the United States would keep its AI lead and framed the technology as a “power maximiser” that could determine global influence.
Xi’s response was comparatively muted, yet experts say China is equally convinced. According to Rebecca Arcesati at the Mercator Institute for China Studies, Beijing sees AI as a key to becoming the leading force, with a long‑term strategy that prioritises open‑weight models that can be freely downloaded and adapted.
China’s rapid release of open‑weight models such as Moonshot AI’s Kimi K3 can match frontier performance while keeping costs low. In contrast, U.S. firms—Anthropic, Google, OpenAI—focus on proprietary, subscription‑based models that dominate current benchmarks, with China’s best recently lagging by a few months.
The global race is not just about speed; it also hinges on economic advantage. China leverages cheaper energy and large‑scale manufacturing, making it the world’s biggest producer of robots and a major player in AI‑powered automation.
Safety perspectives diverge sharply. U.S. officials, citing a lack of existential threats, emphasize self‑regulation and a high‑IQ safety guardrail, whereas Chinese authorities stress “human control” and centralized guidelines to mitigate misuse.
Co‑operation remains uncertain. While past arms‑control agreements dealt with shared risks, AI’s ambiguous threat calculus complicates dialogue. Trump dismissed proposals for a coordinated slowdown, arguing that any restraint might let China close the gap.
Ahead of the summit, U.S. Treasury Secretary Scott Bessent and Chinese Vice‑Premier He Lifeng discussed a potential notification mechanism for critical AI incidents—a step toward limiting unilateral escalation.
As the world watches the exchange, the focus remains on how each power will use AI to reshape economic, geopolitical, and security landscapes—whether through open‑source collaboration or proprietary dominance.
