At Beijing International Airport, it is now mandatory, after landing, to submit hand luggage to the scanner before leaving the arrivals hall.
However, as I approached the carousel with my bags, I saw a sign I had never come across: "Any passenger carrying a drone must remove it from their luggage and declare it." A Chinese security agent pointed directly at the sign to confirm that I wasn't carrying any drones in my suitcase. I shook my head no. I don't usually travel with drones.
That sign wasn't the only new thing I saw on this trip. It was one of many. Traveling to China regularly since the early 1990s, I found this journey completely different for a very specific reason: it was my first visit to the country in the age of artificial intelligence.
As we wrap up nine days in the country, having visited Beijing and Xiong'an, as well as China's tech hubs of Hangzhou, Ningbo, and Shenzhen, it is very clear to me that the AI era is already reshaping the geopolitical and geoeconomic relations between the United States and China in ways that are actually bringing our two countries closer together geopolitically while sharpening our trade divergences.
Like this? Let's go back to that sign about drones at Beijing airport. What does it tell us? It tells us that although the White House and Beijing remain economic and strategic rivals vying for positions of advantage, both realize that there is now an interloper in the marriage between these two superpowers: the hyper-empowered malicious actor.
I'm not talking about defenders of Chinese democracy. I'm talking about agents of chaos: embezzlers, drug traffickers, human traffickers and the kind of hackers who just stole US$6.3 billion (R$33.9 billion) in cryptocurrencies from Binance.
I'm talking about malicious individuals and small teams who can now arm themselves with off-the-shelf commercial drones or AI-powered cyberweapons to single-handedly wield the destructive force of a superpower.
And I'm talking, even more importantly, about North American and Chinese AI models that are demonstrating an increasing ability to spiral out of control and become agents of mass disruption as they become increasingly self-improving.
Both countries are just beginning to come to terms with the shift in how they relate to each other — even as they continue to reflexively take opposite sides in wars ranging from the Gaza Strip to Ukraine to Iran.
Well, it's an obsolete habit that both nations need to unlearn — and quickly — in the age of AI, when global stability will increasingly require mutually assured collaboration between the world's two AI superpowers.
Without a doubt, our definition of disorder is often not the same as the Chinese one. Any regulatory system adopted by Americans must protect democratic values, free expression, and peaceful protest.
China is unlikely to have any of these elements in its regulatory scheme. But there remains an overlap between our interests, which lies in preventing AI from causing violent harm — to infrastructure, to individuals, to countries, and even to our species. We need these ground rules more than ever because AI's next steps involve, quite literally, walking.
I saw a demonstration at a robotics and AI exhibition in Beijing in which a Chinese humanoid robot swapped out its own battery module, housed in its back. It only took a couple of minutes. As I watched that maneuver, I wondered: what happens when he also has autonomous intelligence?
While the need for more collaboration against those trying to bring out the worst in AI technologies may bring Beijing and Washington closer together geopolitically over time, in geoeconomics the opposite is likely to occur.
This is because AI is simultaneously hyper-empowering the strength of China's industry. What I saw on this trip, visiting some of the most advanced companies in the country, is that AI is further intensifying their efficiency and competitiveness.
You can see this in trading statistics. China's trade surplus in 2025 —exports over imports— was an impressive US$1.19 trillion (R$6.5 trillion).
This represented a 20% increase compared to the previous year. This year, it is expected to grow to US$1.3 trillion (R$7.1 trillion) — an imbalance in relation to global economic production that has not been seen by any country since the end of the Second World War.
Trust me: you haven't seen anything yet.
Because while U.S. companies are burning billions investing in AI models so advanced they're humiliating the best mathematicians, Chinese companies are focused on implementing cheap, "good enough" weight models that could make their broad advantage in advanced manufacturing unbeatable.
To put it another way: America is spending immense amounts on AI models that are already too dangerous to launch. China, on the other hand, enters the AI era with such a dominant base of advanced manufacturing facilities that it can reap immediate gains from its “good enough” models.
When you combine this with China's savage capitalism, you get a pace that is genuinely frightening. In the Beijing Robotics Industrial Park, where I saw that robot changing its own battery, another exhibition impressed me even more.
There were two humanoid robots side by side. They were built to compete in the Beijing E-Town Humanoid Robot Half Marathon — 21 kilometers. The first robot runner, who had a gold medal on a ribbon around his neck, won the 2025 race in 2 hours, 40 minutes and 42 seconds.
The second robot runner, also sporting a gold ribbon, won the same race in April in 50 minutes and 26 seconds, surpassing what was then the human world record, 57 minutes and 20 seconds, held by Jacob Kiplimo.
In a single year, Chinese advances have produced a marathon-running robot faster than the fastest human ever recorded. My Chinese guide pointed to the two robots and said, "Last year, this year. Next year—we'll see." We will certainly see. Other than that, folks… nothing to worry about here.
Let me start with good news, what I have absorbed from dozens of conversations and visits to companies in China.
Right now, I would argue that Beijing may actually be more eager to discuss security locks for AI than President Donald Trump, who told the UN in September that Americans "reject any attempt to build a globalist scheme to control" AI. Trump basically says to the world: American models are leading and I will not put reins on them. Full stop.
At the opening session of the 2026 Global Governance Forum in Xiong'an, I had a discussion with Wang Yi, Chinese foreign minister and member of the Politburo, and John Thornton, co-chair of the board of trustees of the Asia Society. I was impressed by Wang's clear message in support of multilateral AI safeguards.
As Wang put it to us: "Leading nations have the ability and responsibility to jointly develop and manage AI effectively, ensuring that its development benefits people. We must not only mitigate risks, but also cooperate to discuss globally applicable rules and standards."
It is clear that Beijing, like Washington, is building tools to enhance weapons and improve every piece of equipment in its domestic surveillance kit.
These officials are under no illusions that Beijing would like to see more locks on America's high-end models because the United States is in the lead. But they will also say that China's law enforcement agencies have become more cooperative with their American counterparts in combating cybercrime, fraud and money laundering — all of which AI is rapidly improving.
So I'm going to go out on a limb here: next year, China and America will be seriously exploring collaborative controls on AI models.
Because as America's models reach the point of training themselves without human direction, and as China's open-weight models are strengthened on that treadmill, I predict that 2027 will be the year that both countries fully understand the incredible opportunity, and threat, that AI poses to humanity.
This collaboration will be complex because, unlike the White House, Xi's government is concerned not only about AI empowering criminals on its turf, but also about political opponents.
Now let's look at how AI is reshaping economic relationships. Every time I travel to China, I come away more impressed than on the previous trip by the stark contrast in its economy.
If China were a person, it would have the torso of an Olympic weightlifter — packed with muscles that represent its export-oriented advanced manufacturing economy — and the legs of an Olympic skier who just ran into a snowdrift, broke his leg and can now only limp on a crutch.
That bottom represents the domestic economy, so weakened by the bursting of its quarter-century-old great housing bubble that many Chinese are holding on tight to every yuan they can.
The frenzied overbuilding of cheap-money apartment buildings in the early 2000s has gotten so crazy that, as the Times' Beijing bureau chief Keith Bradsher wrote in February 2025, China now has "abundant housing for the country's 1.4 billion people and enough empty apartments for another 300 million people."
Put another way, Bradsher said last spring, China has enough empty apartments to house almost everyone in the United States.
Unfortunately, instead of pumping money into the domestic economy through enhanced pension programs, the government is doubling down on subsidies to build its global leadership in advanced manufacturing — and telling Chinese consumers to suck it up.
On my visit to Alibaba's headquarters in Hangzhou, Xiong Xiong, chief of staff to Alibaba co-founder and chairman Joe Tsai, pointed out to me that in Formula 1 racing, you rarely overtake the competition on the straights.
"You win in corners," he said, "because in corners it's not how fast you go, it's how bold you are — how fast you brake and how you respond to my braking — and whether you have the stability to control your car at high speed."
Well, one day we will be able to look back at the years 2019 to 2023 as one of those curves, and China came out of it leading the pack down the stretch. When I visited China before Covid-19, Xiaomi was a smartphone company. When I came back in 2023, it was also an electric car company.
My hosts noted how Xiaomi's self-developed "multi-material performance simulation system" uses AI to quickly discover what the company believes is the ideal alloy formula — out of 10.16 million possibilities — to produce what it called "a perfect combination" of strength and stability for its vehicle bodies.
Global manufacturers used to say you had to be in China "to sell to China." Now they say you have to be in China and have access to their supply chains "to sell to the world."
Do you want to make shoes with electric casters and flashing lights that sing the Chinese national anthem on the left foot and the US national anthem on the right foot? Someone here will have the parts and tools for you to do it in 24 hours.
In Ningbo, I visited Geely's electric vehicle factory. I watched an engineer use the company's latest charging station to power a car from the company's new luxury brand, called Lynk & Co, in just over eight minutes on a charge that will last about 800 kilometers (500 miles).
That's not a typo — it's almost like filling a gas tank. “In the future, we intend to improve even further,” the Geely demo engineer told me laconically. Yes, I thought, I'm sure they will.
In Shenzhen, I visited Huawei Technologies' assembly line for its latest Mate 90 smartphones. Zha Jun, who oversees Huawei's research and development, explained that this line used to have 86 workers per shift. It recently shrunk that number to nine people by turning to AI-enabled image recognition systems.
When I asked what kind of data centers used to run the AI systems, Zha said, "We have no need for a data center. Huawei can get all the AI benefits it needs for this kind of manufacturing improvement using our own open-weight, domain-specific AI models with less than 1 billion parameters."
These are the ones with circuits that are equivalent to the size of a small suitcase. (Parameters are a measure of the connections and weights an AI model uses to process information, and are used as a general approximation for its complexity and learnability.) These small AI models, Zha added, were "good enough" to make big improvements in manufacturing.
Now add one more increasingly important factor: rare earths, which in the age of AI will be more important than ever.
My traveling partner on this journey was Keith Bradsher, who will soon release a book "The Race for Rare Earths: Why China is Winning the Fight for the World's Most Critical Resources."
The book traces how, decade by decade, the US and Europe lost their early lead to China in these vital metals that go into virtually every high-tech invention you can think of over the past 40 years.
In 1964, Deng Xiaoping, who would later become China's supreme leader, visited the world's largest rare earth deposit in northern China. In 1973, he began planning China's rise in rare earth technologies, and in the early 1990s he foresaw the geopolitical power of China's increasing dominance over rare earths, declaring: "The Middle East has oil. China has rare earths."
The West is only now restarting a rare earths industry, and it will take at least five years and probably ten to catch up. Meanwhile, Beijing has exerted its near-total control over the world's supply of crucial rare earths to force the Trump administration and the European Union to repeatedly backtrack on tariffs and to exert strong pressure on Japan not to support Taiwan.
