US President Donald Trump has been blunt about his AI ambitions, as well as about who he considers the US’s main rival to be: China. His Chinese counterpart, Xi Jinping, has not said anything so forthright – but it seems he holds a similar view. “China very much sees it as a race, they very much think China should be the leading force in AI – they see AI as a power maximiser,” says Rebecca Arcesati, from the Mercator Institute for China Studies. You can look beyond the rhetoric for evidence of this: AI has been a cornerstone of China’s technological ambitions for more than a decade, whether Xi emphasises it publicly or not.
For Trump, beating China to AI supremacy outweighs any risks of the technology to humans – something he has described as “a hoax”. China’s foreign ministry warned that “narratives of threat” and confrontation hinder global AI governance, calling for cooperation to ensure it benefits all. It means when the two presidents sit down on Thursday there will be few more important topics than their AI ambitions – and what they say and decide will matter to the whole world. There’s no doubt the US and China are global leaders in AI but who is ahead depends on how success is measured.
American companies lead on so-called frontier models, with the highest-performing systems on most benchmarks coming from firms such as Anthropic, Google and OpenAI. In February 2025, an AI model by China’s DeepSeek “briefly matched the top US model” and as of March 2026, Anthropic’s top model led the pack “by just 2.7%”, according to researchers at Stanford University, external. Part of the US advantage stems from the vast sums being invested in leading AI companies as the frontier labs burn through mountains of cash to build the infrastructure to train new models. The stock market is booming and a large part of the US economy’s expansion relies on the enormous growth – and projected profitability – of AI companies.
So that might partly explain why President Trump sees AI as key to US success. Another factor limiting China’s ability is the sweeping government restrictions placed on US companies that prevent them from selling advanced microchips to Chinese firms – these are crucial for developing cutting-edge models. Despite the chip restrictions, new AI models are coming out almost monthly in both countries and it’s unclear what “winning” the AI race would actually mean, or whether there is a finish line at all. Many companies are racing to develop artificial general intelligence (AGI), a theoretical form of AI capable of learning, reasoning and performing a wide range of tasks at or above the level of humans.
The Chinese AI industry has so far focused not on creating frontier models but releasing what is known as open weight models – downloadable and tweakable in a similar way to open source software. This contrasts with the approach taken by many US firms, which largely focus on proprietary models accessed through paid subscriptions or company platforms. Chinese open weight models are being released at a rapid pace, with benchmarks showing the country’s top models trailing US rivals by only a few months. Models such as Moonshot AI’s Kimi K3 have become popular in China and overseas because they perform close to the frontier models while remaining freely downloadable and adaptable.
Beijing’s focus is more about deploying the technology across its economy, China analyst Lizzi Lee from the Asia Society Policy Institute said. “AI isn’t the moon landing,” she said. “The US may lead at the frontier, while China could still gain enormous economic and geopolitical leverage by making capable AI cheap, open and ubiquitous.” AI is also very energy hungry, and China is very well placed there, Rebecca Arcesati told Tech Life on the BBC World Service. “Chinese AI companies are able to access the electricity they need for a much lower cost and that really is an advantage,” she said.
In robotics – a field of AI that some believe will be world-changing – China has become the world’s biggest manufacturer of robots, with more than two million working in the country’s factories, the International Federation of Robotics estimated. The nation is also the world’s biggest maker of humanoid robots, accounting for more than nine-tenths of global shipments in the first half of the year, according to independent policy institute Chatham House. Chinese AI developers have showed “remarkable resilience and innovation” in spite of the shortage of advanced American chips, analyst Lian Jye Su from tech research firm Omdia said. But US officials and AI firms have also alleged that Chinese developers closed the gap by using distillation – a technique that enables models to learn from more advanced platforms.
China’s foreign ministry said in response that AI development was “fuelled” by the country’s “vision of extensive consultation and joint contribution for shared benefit”. China and the US may be in fierce competition, but their governments’ approaches to safety concerns are markedly different. While some in the industry have recently warned that AI could pose an existential risk to humanity, Chinese researchers seem less concerned about that and are more worried about potential misuse of the technology and cyber war. Responding to concerns over AI, Trump has said the US already has sufficient rules in place and that the only guardrails AI needs are a “high IQ” president.
“We’re leading now over China by a lot, and everyone else, and we’re going to keep it that way,” Trump said this week, adding that the US was not going to “stifle the growth of something that will be bigger than the industrial revolution”. Washington has prioritised innovation and sought to reduce regulation in an effort to “keep American labs ahead”, said information security officer Jayant Dave from cybersecurity firm Check Point Software Technologies. Beijing has taken a different strategy to the US hands-off approach. Xi urged the country’s leaders to keep a close eye on AI’s risks and ensure the technology was always “under human control”.
“China has taken a relatively cautious and centrally-coordinated approach,” introducing guides on data usage and algorithms before many other countries, helping to address many underlying concerns, Dave said. The use of guidelines, though not legally binding, sets expectations while giving developers room to innovate, said National University of Singapore AI ethics lecturer Jonathan Sim. Despite calls for international cooperation, it remains unclear how the AI superpowers plan to address the technology’s potential dangers – if at all. The two countries have previously cooperated on existential threats such as nuclear weapons, but that was based on a shared understanding of the threat, which is not currently the case with AI, Dave said.
“Washington worries about capability escaping control. Beijing worries about capability threatening domestic stability,” he added. Trump has dismissed proposals for an industry-wide slowdown, arguing that it could allow Chinese developers to close the gap. Ahead of the Trump-Xi summit, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng discussed establishing a notification mechanism for serious AI incidents.
While the messaging from Trump on China in recent weeks has been “more antagonistic”, the fact that AI is expected to be on the agenda was a “good start”, said policy analyst Leia Wang from the Carnegie Endowment for International Peace. What we don’t want, he added, was for an “unhealthy version of the race narrative to perpetuate”, potentially escalating an already fierce rivalry in AI. Huge data centres rise at ‘China speed’ to power its AI ambitions Questions mount over what an AI ‘slowdown’ would look like Dramatic insider warnings over AI fall flat with some in Silicon Valley Why are there concerns AI could threaten humanity, and how real are they?
