At first glance, these moves point in opposite directions: Beijing giving its AI companies more access to American technology while limiting foreign access to Chinese technology. But these stories are more connected than they seem. They reveal one of the most important—and poorly understood—developments in China today: the Chinese state is getting more serious about treating AI labs as key elements of China’s overall national power.

That does not mean the government and labs have made a trade or struck a bargain, even though these stories broke around the same time. But as Dan Wang and I wrote in the New York Times after the U.S. government cracked down on Anthropic last month, “We often think of A.I. as a race between the United States and China. Instead we are seeing the emergence of an even more acute form of competition, between the public power of governments and the private power of ambitious companies.”

This competition is unfolding, in different ways, in both China and the United States. The Chinese side of the story deserves more attention than it’s gotten. Chinese labs want access to Nvidia chips that the Chinese government has been loath to give them. AI researchers want to travel overseas but have reportedly seen their trips restricted. Some firms want to relocate abroad or get acquired by foreign companies but have run into resistance from the Chinese government, like when Beijing’s intervention forced Meta to begin unwinding its acquisition of AI agent Manus. Some have contracts with the state or support China’s military modernization, but these days, nearly any conversation with Chinese AI labs also includes bitter complaints about government overreach.

As Dan and I wrote, “Unlike in the United States, in China no one questions who will ultimately win the struggle between state and corporate power.” Even so, the government needs what the labs have to offer and has shown a periodic willingness to make some concessions, while also tightening control over them.

A core question for everyone analyzing China’s AI sector today is defining where the interests of the state and the labs converge and diverge—and how long the current dynamic will last, before a state apparatus that the scholar Dali Yang termed the “Chinese Leviathan” imposes even greater control over these ambitious firms.

Yet a second reality is also on display in these news stories: efforts to ramp up state control over each country’s AI sector and build distinct American and Chinese AI ecosystems are far from eliminating the deep connections between them.

Despite Xi Jinping’s aspirations to self-reliance, China still needs American chips—both licitly, in the case of any approved purchases of H200s, and illicitly, including circumvention of export controls and significant smuggling of the sort alleged in the recent $2.5 billion Supermicro-linked bust. Although the Chinese leadership doesn’t want to acknowledge it, the Huawei alternative cannot yet satisfy China’s sweeping AI ambitions. Beijing’s willingness to consider H200 purchases is further evidence of that gap. And both American labs and the government have reported large-scale Chinese efforts to extract (“distill”) U.S. model capabilities.

The interdependence runs in the other direction as well. Chinese models are increasingly being used by American researchers and companies because they are capable, cheap, and easy to modify—raising concerns in Congress and parts of the U.S. government.

Even much-needed AI regulation is becoming an area of more indirect interaction. The United States has hardly established itself as a coherent model of AI regulation. But Beijing appears to be studying and in some cases drawing lessons from U.S. government actions, including potentially placing limits on overseas access to Chinese models.

These dynamics—the relationship between firms and the state in China, and the fraught but enduring connections between the American and Chinese AI ecosystems—will shape the contours of the AI race for years to come.

***

You can read my and Dan’s piece from June 28 here.

One of the most revealing examples of the transforming U.S. approach to China is something that hasn’t gotten enough attention: how the White House is portraying the visit in videos and photographs. I worked at the National Security Council for several years, so I speak from experience when I say that these images are deliberately chosen and freighted with meaning.

Let’s start with the lead photograph that the White House posted in a series from Trump’s special tour of Zhongnanhai, the Chinese leadership compound:

I think the White House chose this photograph because they wanted to emphasize the fact that Xi gave Trump a personal tour of a part of Zhongnanhai that foreigners rarely visit. But, as a U.S. government product, this photograph is strange for many reasons. Xi dominates the frame, while Trump hovers in the background, slightly out of focus, with a somewhat uneasy or unsettled expression on his face. Xi’s hand gesture makes it look not just like he’s pointing at a tree (or whatever they happened to be looking at) but leading the way. He is large and in charge. Xi’s pose recalls this famous posture of Mao Zedong in much glorifying propaganda:

A photograph potentially chosen to emphasize Xi’s courtesies to Trump actually ends up making the Chinese leader look dominant and diminishes the American president. An odd choice from the White House, to say the least!

The video at this link is even weirder:

The message of the video is supposed to be “American strength,” but the video luxuriates in images of the Chinese military honor guard marching and all the rest of the Chinese-produced pageantry. Again, I think the White House is trying to emphasize that China put on a grand reception for Trump—but the overall effect is, unmistakably, to glorify the Chinese side.

Trump and Xi’s tour of the Temple of Heaven led to another bizarre video, with oversaturated shots of the temple set to a kitschy soundtrack of Chinese traditional music, which could easily be mistaken for a tourist advertisement:

What’s going on here?

Trump, strange as it is to say, is Chinamaxxing.

Chinamaxxing is a viral social media trend that ranges from Gen Z drinking hot water, trying traditional Chinese medicine, and otherwise partaking in traditions associated with Chinese culture, to social media influencers praising sleek Chinese cityscapes as the way of the future. The term is a neologism modeled on “looksmaxxing.” One of the taglines of this trend is a line adapted from Fight Club: “You met me at a very Chinese time in my life.”

There’s a lot more to say about the political and cultural implications of this trend, but it is really remarkable that, after a few days in Beijing, the White House social media account is the latest Chinamaxxer in the mix—not only amping up the president’s trip, which is to be expected, but unmistakably portraying China as a dazzling and imposing peer superpower of the United States.

“This has been an incredible visit. I think a lot of good has come of it,” Trump raved on the second day of his visit. Many other moments reflected this dynamic, like Marco Rubio marveling at the ceiling in the Great Hall of the People and Elon Musk gleefully taking a 360-degree video while awaiting Xi and Trump’s arrival. Even if Rubio has taken a harder-edged tone in some ways—telling NBC News that rebuilding key industries in the United States was something “the Chinese are not going to like…because they want to dominate those industries, but that’s what’s good for the American people”—he has also been clear that Trump sets the agenda: “I execute on the president’s foreign policy,” he said.

As the U.S. debate on China policy continues in the weeks ahead, some supporters of the president’s approach claim that giving China so much “face” is a way to buy time to rebuild U.S. strength, not unlike the now-discarded Chinese approach of “hiding your strength and biding your time.” (It’s “4D chess,” in the words of one friend.) A version of this approach could be part of a coherent and sensible détente strategy, but unfortunately there is little evidence that’s what Trump is actually doing.

If the Trump administration were really running this play, they should be using the period of U.S.-China stability to shore up the sources of U.S. strength rather than undermining them and launching a costly war in the Middle East. The Trump administration is making some notable investments that could serve these goals, including in supply chains for rare earths, though these will take years to come to fruition. But Trump’s war of choice with Iran is wasting money and matériel that would be better conserved to deter China in Asia, and his administration is undermining U.S. strengths by alienating allies and gutting funding for science and technology.

Chinese leaders are not so naïve as to think two days of bonhomie can change the course of history. They are not simply looking for honeyed words from a United States that they see as fundamentally out to get them. Glossy posts by the White House social media account are not going to provide lasting strategic reassurance to Beijing. They know Trump has dramatically reversed himself multiple times in many key policy areas, and he could do the same in his China policy (as he has already during this term—see, for example, the short-lived 145 percent tariffs!).

Xi has made clear that he wants a period of “constructive strategic stability” with the United States (my take here). But as the Chinese scholar Wu Xinbo said after the visit, China’s goal is to “extend our period of strategic stability and win time and space for our development,” facilitating China’s rise. Because China’s leaders believe that time is on their side, they are more than willing to play for time. Their own interests have brought them to this conclusion, not a shift in tone from Washington.

So it remains to be seen whether Trump’s Chinamaxxing moment translates into a meaningfully different approach to China policy for the rest of his second term—or if, like so many viral trends, it’s forgotten as soon as the next big thing comes along.

Meanwhile, the rest of the world is also watching these visuals. They see Trump trying to flatter Xi and promising “historic” deals but coming home with, by most accounts, far less than he’d suggested. These White House visuals underscore the enormous propaganda win that Xi just notched with this trip, which may affect the calculations of countries trying to chart their own paths between Washington and Beijing.

One wildcard in the U.S. debate on China policy will come to the fore this week. The next big event on Xi Jinping’s diplomatic calendar after Trump is a visit by Xi’s closest partner on the world stage: Vladimir Putin.

Putin is slated to visit Beijing on May 19-20. His last visit in 2024 featured some very similar optics to Trump and Xi, including the cheering children that Trump singled out as one of his favorite parts of the pageantry:

After bringing the leaders of the west to China in swift succession—from Canadian Prime Minister Carney to UK Prime Minister Starmer and German Chancellor Merz to the American president—Xi will cap off this stretch with a very different message than the one that we just saw on display during Trump’s visit: a reminder that Beijing’s closest international partners are connected by their shared emphasis on resistance to the United States.

When Xi went to Moscow to see Putin in 2023, a camera caught him speaking to the Russian leader. Gesticulating emphatically, Xi said, “Right now there are changes unseen in a century, and we are the ones driving these changes together.”

Putin replied, “I agree.”

Revealed preferences from Beijing are very different from the talking points its leaders read when the Americans are in town—and, despite Trump’s Chinamaxxing moment, they may ultimately shape the future of U.S.-China relations more profoundly than the opulent optics of one state visit.

One of the core takeaways for me is that Xi Jinping pulled off something Chinese leaders have been working toward for decades: bringing an American president to Beijing as an undisputed peer. He stalemated the United States in last year’s trade war—I’ve been calling this the “stalemate summit.” This past week, Xi used the opulent optics of this visit to put on a show for audiences around the world that had one central message, which is that China and the United States are the two dominant, evenly matched superpowers. There is no going back.

We’re still learning what deals were made in Beijing, but Trump has already made some concerning comments indicating that he intends to treat U.S. support for Taiwan as a bargaining chip with Xi. If he follows through, it’s a clear break with longstanding U.S. policy and is worrying many U.S. allies and national security experts, as Chris Buckley captures well in his article today.

Ahead of the summit, I wrote an essay for the New York Times, “Xi Has Trump Right Where He Wants Him,” on the risks of the current period of strategic stalemate — and how China’s approach to the United States is shaped by the enduring resonance of Mao Zedong’s On Protracted War 论持久战. I’m grateful to CNBC, MSNOW, BBC, and CNN for inviting me on to discuss the piece. I welcome your feedback.

I also joined and to discuss the summit on one of my favorite podcasts, ChinaTalk. Our episode is below:

ChinaTalk

Macartney to Mar-a-Lago

Julian Gewirtz, former Biden administration China official, now at Columbia, joins me to chat about the Xi-Trump visit and all things US-China. Matt Sheehan, senior fellow at the Carnegie Endowment for International Peace, drops by to give his takes on the AI angle…

Read more

3 months ago · 45 likes · Jordan Schneider and Phoebe Chow

During the summit, Xi Jinping introduced a new official formulation for the U.S.-China relationship emphasizing “constructive strategic stability” 中美建设性战略稳定关系. His comments indicated that China’s preferred course is to lock in the current phase of strategic stalemate for at least the remainder of Trump’s term and ideally beyond (未来3年乃至更长时间). Xi presented the search for stability as the central strategic dynamic — acknowledging the reality of U.S.-China competition, as Beijing has done for the past few years, but insisting that it remain bounded and limited. Throughout his remarks, he made clear that China and the United States were peers speaking on an equal footing (平等协商).

“Constructive strategic stability” is Xi’s new tagline for U.S.-China relations at this critical juncture, just as “new model of great power relations” was his formulation for the Sunnylands moment in 2013 when China was still in an indisputably weaker position. Then, as now, the question is whether the United States will decide to accept the Chinese framework, or whether we will get into the game of China claiming the U.S. accepted it and the U.S. saying it didn’t...

Finally, my colleague Zoe Liu and I have pulled together a sterling group to debrief the summit in a Columbia University webinar on Monday 5/18 at 11 a.m. ET, featuring the great Tom Christensen, former Trump 1 NSC China Director Liza Tobin, Washington Post White House bureau chief Isaac Arnsdorf (freshly returned from covering the summit), and China energy expert Erica Downs. You can sign up here: https://columbiauniversity.zoom.us/webinar/register/2717786062854/WN_GAwYJY-nT4Gxh8My6eHBAA#/registration

I’m going to try to use Substack more regularly in the period ahead. Feel free to send over suggestions for topics or questions to tackle!

In April, Trump played what he believed was his best card, sky-high 145 percent tariffs, but quickly dropped them. China managed to expand its global exports and found new markets for many of the goods that it previously sold to the U.S. By contrast, as Beijing weaponized its chokepoint over the critical minerals essential to modern industry, Xi discovered that his best card was even stronger than he expected.

By agreeing to delay its new critical minerals policy by a year at the recent Trump-Xi meeting—in effect, leaving the weapon on the table and promising not to use it as long as the U.S. behaves—Beijing has managed to entrap Washington in an open-ended negotiation in which both sides are acting as if China has the upper hand.

The U.S. needs a China strategy fit for purpose at a time when China has demonstrated an increased willingness to weaponize its leverage.

China will be watching several areas closely. Will the U.S. make meaningful investments at home? How will Trump’s strained relationships with longstanding U.S. allies evolve? But Beijing is especially focused on whether the U.S. uses this break from brinksmanship to devise credible, hard-hitting escalation options that could be employed against China in the future. Beijing fears that Washington will deploy the threat of powerful chokepoints and retaliatory measures of its own—the kind of U.S. leverage that would make Beijing hesitate before wielding its critical minerals weapon again.

Our piece describes how the United States can win the economic war with China. But Trump appears unlikely to take that path.

Read it here: How to Win the Economic War with China.

By this point, everyone who follows China has probably heard of Dan’s framework—China as the “engineering state,” the United States as the “lawyerly society”—so I pushed him on a few key points:

• Ten years ago, the Made in China 2025 plan set ambitious goals for the Chinese manufacturing sector. It’s now 2025. Did the plan work?

• How should we make sense of the ironic reality that the Chinese economy has a small percentage of very high performing firms and a massive percentage of firms that we might charitably call deadweight?

• Dan critiques the United States as a “lawyerly society,” in a vein similar to Ezra Klein and Derek Thompson’s Abundance. But many of us are very worried about threats to the rule of the law in the United States today. Does U.S. democratic backsliding mean we need to update or rethink aspects of Breakneck’s framework?

• Is the United States really a “lawyerly society” first and foremost, or do we need to focus more on finance and Wall Street run amok to understand how we got here?

• What is the right end-state goal for U.S.-China rivalry?

Check it out:

First, I’ve argued that the Trump administration is undermining the United States’ competitive position in ways that are more profound and likely more enduring than many people realize. Amid all the ups-and-downs and back-and-forth, it can be hard to see the bigger picture. But I believe the risks to the United States’s sources of strength at home and abroad are deeper than many people realize. In an essay called “How China Wins” in the July/August issue of Foreign Affairs, I laid out my view:

U.S. President Donald Trump and Chinese leader Xi Jinping are ushering in an era of what can be called “mercenary multipolarity”… Trump and Xi are turning U.S.-Chinese competition into a story of two self-interested, domineering superpowers looking to squeeze countries around the world—and each other—for whatever they can get. This dramatic shift plays into China’s hands and undermines core U.S. strengths in the long-term competition over the future international order.

Subscribe now

I elaborated on this further in an interview with The Wire China:

Beijing’s assessment right now is that the United States is dismantling, fairly systematically, the sources of its strength. The United States, in their view, is dismantling its alliance relationships and alienating much of the world. It is dismantling aspects of the U.S. science and technology ecosystem, cutting funding to some of our great universities, and making it very unappealing, if not outright impossible, for foreign talent to come do research in those universities. And it is eliminating arms of U.S. influence around the world, from USAID to Voice of America. China’s view is that the United States is, in a sense, unilaterally disarming.

Second, I’ve argued that common depictions of a U.S.-China “AI arms race” miss key aspects of China’s drive toward advanced AI and the implications for geopolitics. Last week, I wrote in an essay for the Financial Times:

Both countries are investing heavily in the technology. But, outwardly at least, they appear to be pursuing different goals. While US leaders prioritise developing the most intelligent models possible, Chinese policymakers are most focused on AI’s widespread application. To use Silicon Valley parlance, the Chinese Communist party seems far less “AGI-pilled” than their American counterparts…

What this means is that even if the US and China are running what look like different races in AI, direct competition remains intense. Whether Xi doubts AGI, feels constrained, or is concealing his intentions, there is a profound risk on both sides of “technological surprise” — when a rival gains an unexpected technological capability, it can escalate the risk of conflict…

These scenarios drive home a key strategic insight: in isolation, both the intensification of US-China rivalry and the explosion of AI breakthroughs are world-changing. Together, their simultaneity and interconnectedness creates unprecedented risk.

Third, I’ve argued that China’s global strategy is more confident, not less, in the Trump era — and Beijing has adapted not by ramping up a “charm offensive,” but rather by increasing pressure on U.S. allies that China sees as weakened due to Trump’s disdain for alliances. On the day of the EU-China Summit last month, I wrote for Foreign Policy:

China’s assessment of European power is the most important factor shaping its current diplomatic approach. The Trump administration’s recent concessions the U.S.-China trade standoff have confirmed China’s confidence in a hard-edged strategy of pressure and economic coercion—toward Europe as well as the United States and its other allies…

China remains especially diplomatically active in neighboring Southeast Asia and parts of the global south and will continue to take advantage of opportunities presented by the Trump administration. But rather than needing to launch a new systematic global bid for leadership, it feels that position is currently being handed to it. So the United States should take no comfort from the lack of a Chinese charm offensive, because what it reveals is something much deeper: Beijing’s profound confidence that it can exploit this era of “mercenary multipolarity” to its enduring advantage, at the cost of both European and American interests.

I’ll be using this Substack to discuss these ideas and many others on China, U.S.-China relations, U.S. strategy, tech, supply chains, and more.

Finally, on the theme of building new things—and thanks to ’s urging to include some art/culture in every post—here’s an amazing photograph by Marc Riboud of the construction of the first bridge over the Yangtze River, taken near Wuhan in 1957:

Marc Riboud, Forbidden City Under the Snow, Beijing, 1957