Subscribe now

On to the post… –Graham

The Chinese company Moonshot AI released the weights for its much-discussed new Kimi K3 large language model today, as it promised when the model debuted exclusively through their platform on July 16. This means anyone with the (considerable) requisite computing resources can now download and run the model without interacting with its maker. The 1.56TB package is now available on Hugging Face.

The license Moonshot attached to the model restricts one key type of use, however. If you wish to provide the Kimi K3 model as a cloud service to resell, and if you have revenue over $20 million USD per year, you’ll need to “enter into a separate agreement with Moonshot AI.” This $20 million annual revenue threshold is comically low; almost all cloud platforms that resell models already trigger this condition. In effect, any inference provider—from Amazon Web Services down to GPU-first “neoclouds”—will have to enter into a commercial relationship with Moonshot to serve up Kimi K3. It’s a good bet that Moonshot will get some revenue sharing out of these deals. Companies or researchers that want to use Kimi K3 internally should still be able to do so without paying a dime. But if they want to sell the model as a service, they’ll have to make a deal.

Given how expensive it is to train a model of Kimi K3’s size and capabilities, attempting to make some money is understandable, as Moonshot stares down a not-so-distant future of becoming a public company. However, at a time when the US government is debating various ways to potentially block or discourage US use of Chinese-built open-weights models, this desire for a new revenue source also introduces a new source of risk.

Making Revenue Sharing Explicit

Kimi’s new license provision looks like an attempt to walk a tightrope.1 Engineering and training large language models is expensive. The natural (though far from proven) business model to recoup those costs and hopefully make a profit is to sell access to them. Yet Chinese labs have released many of their most capable models openly, allowing companies and well-resourced individuals to run them for free. The only way they’ve been making money so far is serving their models behind their own APIs, which has limited reach. This means other companies are capturing some of the revenue that the open model makers might theoretically have collected.

On the other hand, openness is good for visibility and feedback. For instance, practically no one in the United States had heard of Moonshot AI, founded in 2023, until a gradually growing cohort of users found earlier Kimi models effective and cost-efficient for coding tasks over the last year or so. The first generation Kimi model, Kimi 1, was actually a closed model. The reaction and traction was so poor that the company decided to open weights for future generations. This month, with that growing visibility, Kimi’s K3 release landed on the front page of the New York Times.

To the extent that prominence drives direct business and investor interest, openness helped Moonshot level up quickly. Moreover, Moonshot has limited computational resources to serve model users; it even had to pause new sign-ups amidst the rush of attention around the K3 launch. And some users will not want to send their sensitive data outside their own company’s trusted computing infrastructure, let alone to systems controlled by a Chinese company.

So while China’s cloud-computing leader AliCloud has explored a business model where the most advanced version of their Qwen model was available exclusively from Alibaba’s cloud, hopefully catching some revenue and getting users used to their platform, Moonshot has tried something different. The company almost surely needs other providers to meet demand, and they can cut deals to bring home some of the revenue.

Changing the Landscape of Regulatory Risk

This sounds like a good way to retain the benefits of open-weights release while pulling in some additional revenue to make the whole business more sustainable. Yet this license changes the nature of Moonshot AI’s legal relationship with inference providers around the world.

Previously, the providers could download the model, get it running on their infrastructure, and start selling tokens. No need for any explicit deal with the Chinese firm behind the model weights. Peter Harrell, a former Biden administration official and an expert in export and transaction controls, has argued that it’s unclear whether courts would allow the government to prohibit this pure form of open-weights model use. Considering various potential policy tools, he suggests that downloading the open models could be deemed an exempt transfer of “information,” rather than a controlled “service” or other transaction.

But if a US company needs a contract with Moonshot to provide the inference tokens that Kimi K3 generates, the picture looks different. Some of the policy tools US officials and others have debated as potential levers to restrict Chinese open model use would more clearly apply. For example, if the government decided to use the so-called ICTS rule, thus far a little-used mechanism by which the Department of Commerce can block “transactions” with “foreign adversary” entities that it deems to pose national security risks, it’s harder to argue there’s no transacting going on. If they put Moonshot on a sanctions list of entities US persons cannot legally do business with, then the contract to provide inference is clearly an instance of doing business. The commercial agreements this new license is asking for (and getting) from US companies will avail Moonshot to American jurisprudence on a whole new level.

This isn’t an argument about whether such restrictions are a good idea, but an analysis of the policy implications this new license introduces. As Moonshot seeks to generate more revenue from its work while still holding onto the halo effect of being “open,” it seems to have walked out onto a tightrope with a new balancing act: With revenue comes regulability.

Share

Subscribe now

About Here It Comes

Here it Comes is (with occasional coauthors!) written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

There is a great deal of debate about how far “ahead” the United States might be in AI, and usually the “lead” is measured in months. While I stand by my earlier argument that the notion of an “AI race” between countries is more confusing than helpful except when it’s carefully defined, we can certainly say that there are differences in utility among various LLM products.

Let’s assume there is a set of tasks that a skilled user (whether an individual, a company, a government, a criminal group, whoever) can today get done with a US LLM product but not with a Chinese one. Take this as a functional definition of a “US lead” without assuming a unilinear development path or some particular destination. Sounds like a safe assumption, and one completely aligned with the consensus idea that the United States is ahead.

Now, should we not at the same time assume there is a set of tasks the same user can get done with a Chinese LLM product but not with a US one? At minimum, we could start with any task that requires relatively high performance on some benchmark AND the ability to run the model on the user’s preferred infrastructure without permission from the developer. In this way, we can understand a “Chinese lead” to exist, at minimum, where openness enables certain uses.

So there you have it. If we understand AI capability as a set of user-task pairs facilitated by a given system, there is a complex landscape of exclusive capabilities. It also means that a lead, or more properly an exclusive capability set, looks different to different users.

A US lead exists, if you like, but only if you recognize that a Chinese lead also exists.1 The other option, of course, is to reject the idea of being ahead or behind in raw technology and focus on specific actual changes in the world.

Share

Let me know what you think. I’ve got covid brain this weekend (it’s still out there, folks!) so may not be my sharpest.

Before I sign off, a little shameless self-promotion:

• I was lucky enough to share some thoughts for an NYT article by Meaghan Tobin and Cade Metz headlined “China’s Latest A.I. Breakthrough Threatens America’s Lead.” I think it’s an important piece, and it got me in the brain space of thinking about leads again. To me my most important intervention is that, contrary to an often-heard argument that everything China builds in AI is effectively stolen, “The Chinese models are not excellent only because of distillation. There is real innovation going on.”

• I was also lucky to share early thoughts on the announcement of the World AI Cooperation Organization launch in Shanghai with the WSJ, in Jonathan Cheng and Katrina Northrop’s article headlined “China’s Xi Touts Open-Source AI and Takes a Swipe at U.S. Dominance.” Some further details on WAICO (namely the signatories) have emerged since we spoke, but to the extent WAICO is interested in AI governance, my observation stands: “Governing AI risks globally ‘is by definition going to need the U.S. and China both involved,’ said Webster, who focuses on Chinese tech policy. Otherwise, he said, ‘it can’t be a comprehensive global solution.’”

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

What does the Reuters report, based on three “people familiar with the discussions,” suggest about the future of open-weight AI releases out of China? A brief discussion and notes on what to watch…

What’s at stake if China limits open-weight releases?

A Chinese government action to limit the open-weight release of LLMs would be a revealing indicator of what officials are worried about when it comes to the current paradigm of AI.

Releasing a model’s weights means anyone can download it, use it on infrastructure they control without the model maker’s supervision, and adapt it to build new systems. Some people believe open-weight releases already lead to a dangerous diffusion of capabilities in cybersecurity, biology, and other realms. China’s top developers have generally gone for open-weight releases, leading to a diverse and relatively cutting-edge ecosystem of open models out of China that are used around the world. I described this landscape with Stanford colleagues in December, and China’s open offerings still dominate.

The Chinese government has regulatory power at the ready to limit open releases, so the absence of such limits suggests officials are not concerned enough about open-weight models leading to security risks that they deem it necessary to block them.2

This has been a much-watched area. In April, when DeepSeek finally released its long-awaited V4 follow-up model, Alibaba had recently kept its top-of-the-line model proprietary for the first time. DeepSeek’s open release, another by Moonshot AI, and subsequent releases from others including Z.ai suggested Ali’s decision was a matter of company strategy, not government influence—but some of us had wondered it Ali would be a bellwether.

The Chinese government could still decide open-weight releases are too risky or undermine national competitiveness in some way. Something like what Reuters reported would make sense as a first step in such an intervention, but it’s not at all conclusive. Regardless, since Chinese labs provide practically all of the up-to-date open LLMs in the world today, users of open models anywhere stand to be affected.

What did Reuters report, and what should we watch for?

The article, by Fanny Potkin, covers several overlapping themes. I’m focusing here on those relevant to open-weight limits. Here are the key passages:

• “At the meetings, led by China’s Ministry of Commerce, participants discussed putting limits on the most advanced AI models — both closed-source and more open versions, according to two of the sources.”

  • COMMENT: Once a model’s weights are released, it is not possible to limit its use on someone else’s server. In the story’s framing, model makers could be talking about post-training models before open release to limit them in some way or about withholding the most advanced versions from open release (something Alibaba has done already, apparently due to business reasons).

  • WATCH FOR: (1) Release of closed-only models from labs that previously put out their top product openly; (2) Indications that open-weight models have incorporated government-required limits.

• “The scope of the potential restrictions is still being discussed, two sources said, adding that they may only apply to future models. It was not immediately clear when or even if they would come into force.”

  • COMMENT: There would be no practical effect to post-facto withdrawal of an open-weight release, because there would be copies online, but it’s possible firms would pull them down from their main release platforms as a matter of formal compliance.

Of course, a policy move could come with an explicit announcement and not the reading of cryptic signs from unnamed sources. The Chinese-hosted World AI Conference is coming up again this month in Shanghai, which may be a venue to signal any new approach. Stay tuned.

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

If you will indulge me, in the tradition of my occasional series in this space “Didn’t See That Coming” (DSTC), here’s a 20-track DSTC megamix of stuff I would not have anticipated 20 years ago. Plus: a few things I think I would have told you I did think were coming that did pan out.

This is just for fun, but it’s also something I’ve been thinking about when I speak with students and younger folks about their future plans. One of my most consistent messages is that, at least for a path like mine, you couldn’t plan it out on purpose if you tried.

Didn’t See That Coming: Megamix

• #20: A D.C. Metro line to Dulles airport. The cars were old, the stations were epic brutalist displays that felt poised for use as fallout shelters, and adding Dulles to the map seemed pie-in-the-sky.

• #19: Two White House terms for the junior senator from Illinois and Northwestern ’06 commencement speaker. I liked his clarity against the Iraq War, and it was a great speech, but at this point in 2006 I wouldn’t have even bet on Barack Obama in a primary.

• #18: President Donald Trump. Obviously. Holy hell.

• #17: A cumulative ~four years spent in Beijing, but mere weeks in Japan. In 2006 I finished a senior thesis on Sino-Japanese relations, and I decided I would start learning Mandarin and move to Beijing for the year leading up to the 2008 Olympics. But I’d spent time in Japan and had passable Japanese, and I would not have guessed that China would so dominate my professional life over the next 20 years.

• #16: Trader Joe’s on 14th St. NW. I’m not even sure I had ever seen a Trader Joe’s. The first visit I remember might have been 2007 in Palo Alto. But by the end of that year I knew 14th St. NW and it continues to amaze.

• #15: China’s centrality to advanced technology supply chains. I knew a lot of electronics were made there, but I didn’t even know how to ask whether this might lead to whole categories of to-be-developed technology that is really only available from China.

• #14: China’s sustained turn toward repression. It’s not as if I thought in 2006 that the Internet would cause freedom in China. My college thesis had analyzed tech use both for mobilization of protest in China as well as party-state efforts to tamp it down. Yet China appeared to be trying to open, and it would be two or three years of before the progressively less free course we still see today would take hold. Stuff changed and it will again.

• #13: The blogosphere is dead. Long live the newslettersphere! I’ve had personal “Web sites” and blogs since I was a preteen. In 2006, blogs were king (though I had argued news orgs should innovate further given their resources). The death of the blogosphere at the hands of Twitter and its reinvention through Substack and others would have blown my mind.

• #12: The attention-abusing intrusiveness of social media. In June 2006 Facebook still seemed exciting and fun. That September, when the introduction of the “News Feed,” I would realize this was on a path many of us wouldn’t enjoy and that our sense of intimacy and privacy was an illusion.

• #11: I would learn a ton about the UN Convention on the Law of the Sea by a decade later, and then forget most if it as another decade passed.

• #10: Teaching grad students at Stanford. I knew I wanted to go back to school—for international policy or Asian studies or something. I don’t think I anticipated my future would involve teaching real adults at a serious and fancy university.

• #9: I like sour cream now. More specifically, Mexican crema. This happened when a roommate in Seattle introduced his nachos method.

• #8: I really liked graduate social science statistics. In high school, math was interesting but a bit of a drag. A required stats course as a journalism major sparked no interest. Prof. Chris Adolph’s year-long sequence at University of Washington permanently upgraded my BS detector and, later, made it possible to kind of understand machine learning.

• #7: I went for a PhD in political science. I didn’t have much understanding of the discipline, and I wouldn’t understand that it was a potential path forward until studying for my master’s.

• #6: I quit the PhD. Pretty sure 2006 me would have assumed I’d finish what I started, but turns out I was willing to change course when it turned out I didn’t see the path I’d hoped for. Good.

• #5: A string of half marathons. Circa 10 years later I would start doing about one a year. I don’t think I’d ever run more than 7 miles by 2006, nor did I want to. Hope I can continue.

• #4: AirPods. Absolutely wild.

• #3: Friendships that snake around the world. You meet people in one chapter, and they show up somewhere else. I think when you’re 22 this is incomprehensible.

• #2: The things you can remember. The nervous sensation of being a Colorado kid in a think tank meeting with people who seem so important, it turns out, can waft through your awareness when you are one of those people (not important) going about a regular day at a think tank table.

• #1: The things you don’t remember. As a young journalist I would remember what people said verbatim, without trying, for a couple of hours. Where did this tool go? And what else didn’t I see coming that there’s no way to recall?

Saw that coming (please clap).

• The iPhone. When Steve Jobs announced the iPod in 2001, I was disappointed that after the build-up it was only a music player. I already thought it should be a phone, and it took them six more years to catch up to my vision.

• Living in Northern California. I was at least going to give it a try one day.

• Not really picking a career lane. By 2006 I was figuring out that journalism alone wouldn’t do it for me, nor would straight policy work, nor traditional academics. I think I’d have accurately guessed some kind of blended path would continue to unfold.

OK, back to regular programming next time. If you want, reply or share this post with your favorites, your own shockers, or the times you saw what I couldn’t see.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

There’s no sense in picking on particular journalists or pundits. The shorthand of a “global AI race” is so omnipresent that it’s a natural crutch. (Brevity is a virtue in headlines and broadcast scripts.) We certainly could pick on some specific actors who have intentionally pushed this frame, but that’s not my purpose here. As US President Donald Trump prepares to travel to Beijing for a state visit and the “race” rhetoric soaks the media, I want to outline why I think it confuses everyone about what is actually going on.

Here, in brief, is my answer, in three parts.

1. There is no comprehensive ‘AI race,’ because AI is not one thing.

AI can mean facial recognition, social media feed algorithms, deepfake video production, LLM-based chatbots, autonomous robotic systems, and more. A race for such a thing can imply competition for technical excellence, engineering toward practical applications, actual deployment, revenue, or geopolitical advantage. If people want to talk about being ahead in an AI race, they need to specify which one. Which technology, and what aspect of its role in the world?

The idea of nation-vs-nation AI competition between the United States and China has been present in public policy debates at least since the Chinese government’s New Generation AI Development Plan (AIDP) in 2017 declared:

[B]y 2030, China’s AI theories, technologies, and applications should achieve world-leading levels, making China the world’s primary AI innovation center, achieving visible results in intelligent economy and intelligent society applications, and laying an important foundation for becoming a leading innovation-style nation and an economic power.1

US commentators reacted to this ambition with alarm, though some also pointed out that the AIDP seemed to react to Obama administration plans from the year before. (This observation was later confirmed to me in a US-China dialogue when one of the people who worked on China’s plan directly said the US plans helped drive China’s own planning.) If nothing else, we might observe a tit-for-tat dynamic in AI competition rhetoric.

Yet in 2017, AI was not large language models. It meant various forms of machine learning, largely of the “deep learning” paradigm, focusing on pattern matching. Facial recognition and computer vision for medical diagnosis were rising. A DeepMind system had beat the Go champion Lee Sodol the pervious year. The AIDP and other future-oriented documents contemplated all kinds of other things, but no one knew what would emerge.

Fundamentally, if you’re going to talk coherently about “winning” a race, or a competition or battle or war or whatever metaphor, you’ve got to know what you’re talking about. The “AI race” idea, unless precisely narrowed, leaves it unclear how to judge success or failure. It’s evocative, but empty.

2. A ‘race’ has a finish line. Unless you specify one, the term frustrates careful thinking.

Even if you define an “AI race” in terms of being “ahead” or “behind” on some specific metric, say the best performance of any system in each country on a chosen benchmark, you still have the problem of the race metaphor implying a destination. If there’s no finish line specified, it will be impossible to judge the truth value of any speculation about who will win, or who is winning. The lesson of the tortoise and the hare is that being “ahead” doesn’t mean anything until the finish line is reached; if no destination is specified, you can’t measure victory.

Defining a specific goal is a good, serviceable way to use the evocative and sometimes accurate language of “racing” to develop something. I have no beef with anyone talking about a race between firms to be first to market with a product that reliably performs a given task, or to be the first to make such a thing profitable. If you talk about a race to develop a system that aces a particular benchmark, that too is specific. You can talk about a race to build an autonomous driving company with 10 million vehicles operating per month. Yet these are not “the AI race”; they are “a race to _____.” If anything, the simple exercise of asking about specific goals reveals that there are many, many races in “AI”—and that only some of them make sense as a nation-vs-nation contest.

Now in many cases the implied destination is artificial general intelligence (AGI). Yet the concept of AGI has always been speculative and diffuse. Some ideas of the thing have clearly been achieved. Others are so far outside of what we witness today that it’s hard to see any “race track” ahead. A race to AGI simply is not a specific thing unless a specific definition is agreed.

3. What’s good for a US company is not necessarily good for America.

Suppose we agree that one or more labs, or constellations of organizations, are in a race to achieve a specific outcome. If a well-defined finish line is achieved first in a given country, does it mean that country has won? Not necessarily.

Let’s say Waymo’s deployment of highly autonomous ride-share vehicles that do quite well in a lot of different contexts has just reached a well-defined finish line first. Waymo, or Alphabet, could naturally be said to have won this specific race. Some of the many companies Waymo beats in this hypothetical are Chinese.2 Has America won and China lost? The US government doesn’t own Waymo, nor do the American people. Some shareholders, who are spread around the world, might gain. Some riders enjoy the service. If it is wildly successful, already-struggling gig workers likely will not enjoy the increased squeeze on their finances. And Chinese life proceeds largely unaffected.

How about a more specific and complex example. Say that we were aliens with advanced knowledge of human computer systems monitoring for the moment an LLM system (and the specialists at the company that made it) could discover the specific OpenBSD vulnerability Anthropic’s Mythos reportedly uncovered. And say no one in a cyber offense research group has been able to secretly use other available systems to discover this secret so far. Anthropic has just “won”! And the company is making some efforts to spread the benefits of this achievement among US actors and not Chinese ones (though Chinese actors also win when they implement the OpenBSD patch). Anthropic appears to be trying to turn its achievement into a win for the United States, and/or it is trying to encourage the perception that this is the case. Here you can make a coherent argument that the United States may be winning marginal cybersecurity advantages, one model at a time, while China is denied this edge. But that doesn’t tell us anything about the overall situation with “AI.” The problem of many forms of AI (#1 above) and the general absence of specific finish lines (#2) remains

Again with the AGI: Many people who feel urgency about a US-China “AI race” will tell you they are concerned about self-improving, potentially catastrophically capable systems. Sober-minded concerns about systems that may be deployed in ways that lead to terrible outcomes are blended with singularity, Terminator, Matrix, and human-extinction imaginaries. The fear of unknown capability leads to the fear that an adversary might get to the unknown first, holding our team—whoever that is—at peril. So shouldn’t we throw out all these objections to the “race” metaphor and unleash all US actors to blunder forward into the unknown lest someone in China get somewhere first? To this I simply ask: If the fear is of unknowable, unintended, uncontrollable consequences, doesn’t it take a healthy dose of hubris to imagine that just because a system is made by OpenAI or Anthropic, or xAI, therefore it will naturally operate to the US benefit? Which benefit specifically, and for which US citizens? We have longstanding imperfect institutions designed to align the US government with the plural interests of the US people, and those are under extraordinary strain. What precisely should an idealized, “aligned” such system conform to?

Conclusion

Talking about a single, generalized “AI race” between the United States and China distracts from the multiplicity of endeavors under the AI banner. It implies a clear finish line when most often the discussion at hand has specified no such thing. And it inaccurately assumes the national identity of an actor determines which national interests their work will benefit.

So if you find yourself reaching for the “AI race” in a headline or a commentary, ask yourself if you can be more specific. Ask in what realm competitors might be operating, ask “race to where,” and ask who benefits if someone gets there first.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

1. Is China turning the corner away from a norm where many of the most advanced LLMs are released openly and available to the world for download, deployment, and adaptation?

2. Will Chinese open-weight models continue to be fast followers of US closed models in terms of capability?

Today I think it’s clear that the answer to the first question is no: Rumors of the death of Chinese open-weight LLM advancement have been greatly exaggerated.

As for the second question, I think we need more data from the community to answer fully, but at minimum, we don’t seem to be observing a sudden stall in Chinese model metrics you might expect if chip controls were biting hard.

Busy day

In April 2026, Chinese open-weight AI development is alive and well.

• A few hours ago DeepSeek released its much-awaited V4 model, sharing it under the MIT License, effectively as open as it goes. According to the company, the DeepSeek-V4-Pro variant at maximum effort now leads the open-weight ecosystem and rivals closed leaders on a number of metrics.

• Not to be outdone, Moonshot AI also in the last few hours released its latest offering, Kimi K2.6, on a modified MIT license — also reporting incremental gains in performance. Anecdotal conversations suggest Kimi models have been popular with agentic and code-assist users looking for less expensive options compared with the closed US leaders.

• Even Alibaba is still in the open-weight conversation, releasing a version of its Qwen 3.6 also in the last few hours.

We’ll see soon enough how all these perform in various workflows, and how they stack up in price-to-performance calculations.

The open-weight question

There truly was reason to suspect that the pattern of Chinese labs releasing their highest-performing models could be coming to an end. Alibaba’s move to keep its new marquee offering, Qwen 3.6 Plus, closed was a break for China’s cloud leader, which was releasing open models even before DeepSeek made it cool.1 This could have been a harbinger of a shift in business models or backroom regulatory pressure, or—as seems to be the case—it could be one company’s change in strategy.

There were vibes-based reasons to wonder whether China’s government, which generally tries to keep the domestic cyberspace environment well supervised, might now view highly capable open-weight models as an unacceptable risk. After much-publicized adoption of the OpenClaw agentic system, which can behave unpredictably, authorities reportedly instructed government officials not to raise their lobsters on work machines. The publicity around Anthropic’s Mythos model and its purported transformative cybersecurity capabilities fuels worries around the world that increasingly advanced models could empower malicious actors.

Did hobbling China work?

As I say, I think the jury is still out about whether chip controls are biting. Proponents of the Biden-era policy, who would generally have favored increasingly tight controls, have been alarmed by the Trump administration’s approach. It’s not clear just how much cutting-edge GPUs are getting through. And it’s not clear how much of the training for these models was done domestically on US-controlled hardware vs. domestically on Huawei chips vs. internationally in data centers not covered by export controls.

I started to hear that DeepSeek V4’s release was imminent at the end of 2025. Then I heard it was coming by the end of the month in at least two previous months this year. This time, it’s out, but whatever caused the delays may indeed have to do with computational power. Judging the implications here on a US-China strategic lens will require better understanding of the eventual product versus what the US labs are able to put out in this interval with their unencumbered access to NVIDIA’s latest.

Agree? Disagree? Let me know.

Share

Subscribe now

Programming note: In February I promised monthly posts; underestimating the mental load of travel and teaching, I did not deliver in March. So I owe you all—and myself—another. Stand by!

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

The announcement is written in the style of a cybersecurity firm that has uncovered a major hacking network and is here to dutifully report current events while, by the way, offering cyber defense services. In other words, there is the news and then there is the message. Here, there is also the context of a company at odds with the US administration.

The news is that Anthropic says its “terms of service and regional access restrictions” were violated in service of “distillation” efforts, where the outputs of one model are used in the training of another.1 That’s a fair complaint for a company that has strong views on proper, safe use of its technology and is in a very competitive market. The company provides interesting and informative details about the types and volume of the detected activity. has a good read this morning about what it all means and doesn’t mean from a technical perspective.

For me the message is more interesting. The title calls the efforts “attacks,” though there is no indication of service disruption or data breach. The moral valence of the language— “attacks,” “illicit,” “fraudulent,” “threat”—aligns with the company’s broader stance that Chinese AI labs must be constrained lest their capabilities empower a geopolitical rival and authoritarian rights abuses.

The context of Anthropic’s policy advocacy with the US government is distinctive and important for the company. During my conversations in Washington last week with people in government, research, and corporate roles, the company’s voice kept coming up, favorably and unfavorably. The current stakes are especially high. Axios reported yesterday Pentagon officials have called in CEO Dario Amodei with a $200 billion contract in question. “This is not a friendly meeting,” a DOD source reportedly said. “This is a sh*t-or-get-off-the-pot meeting.” Earlier, Secretary of Defense Pete Hegseth reportedly threatened to declare Anthropic a supply chain risk.

So what is Anthropic’s message, amidst these high stakes?

• Under “why distillation matters,” the company argues “national security risks” result from lack of safeguards in models produced through the disclosed type of activity. They invoke “bioweapons” or “malicious cyber attacks.” Anthropic earlier announced that it had attributed use of Claude Code for cyber offense to a “Chinese state-sponsored group.”

• They raise the possibility of “enabling authoritarian governments to deploy frontier AI for offensive cyber operations, disinformation campaigns, and mass surveillance.” More national security plus some human rights.

• If models are “open-sourced,” they write, “this risk multiplies.” Anthropic wants advanced models to be closed.

• Amidst distillation, “the apparently rapid advancements made by these labs are incorrectly taken as evidence that export controls are ineffective and able to be circumvented by innovation.” Anthropic is widely seen as a key advocate for export controls targeting China, and they seek to refute arguments that the controls don’t work.

What their message isn’t:

• Anthropic doesn’t emphasize the potential rise of competitors if distillation contributes to useful products in the same market.

• They do not emphasize theft of intellectual property or industrial secrets. (Perhaps this is because AI companies are not exactly morally pure on the IP front?)

In sum, the message is that Anthropic cares about national security and human rights, and, in service of this, wants the most capable models closed-source while maintaining export controls designed to prevent Chinese labs from meeting or exceeding US labs’ capabilities. And they portray themselves, accurately in my perception of the field, as a leader in this worldview—talking about “intelligence sharing” with other labs.

I find myself asking here about the mixture of a principled stance vs. strategic communication. On the principled side: Whether one agrees or not with the principles—and many disagree from many angles—there’s a lot of continuity here. Anthropic and Amodei have long been vocal about their view of safety and the geopolitical valences of AI. On the strategic side, it’s reasonable to ask if the national security and “attack” language was tuned or timed to ease the Pentagon discussions. At the same time, in “what their message isn’t” above, we see a foregone opportunity to pander to a skeptical administration where influential factions care about US firms leading AI globally. The company simply doesn’t make this about its own, or the country’s, bottom line—a frame that has been useful for others projecting alignment with the US government against China.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Monthly: Back when this publication was U.S.–China Week, from 2015–18, the weekly rhythm was designed to keep me current and prompt distilled thinking through writing. This time, monthly will be enough.

Keep it brief: With the popularization of newsletters/the return of the blog, there is too much quality material. Reading through the excellent offerings, I see writers unleashed. The good news is we have more, better-informed voices digging and thinking than before the Twitter diaspora and the Substack hustle—maybe better than the blogosphere before tweets shrank our time horizons and attention spans. The bad news is that writers unleashed are often also writers un-edited (or LLM-edited) and undisciplined in terms of length. Brevity will take longer, but it will hopefully make things easier for readers and require me to think more clearly. My target will be 500–800 words, with a hard stop at 1,000.

While I’m making commitments, I might as well say that LLM products do not and will not touch my prose. I absolutely use them to surface source material, but just as with a search engine, the reading and characterizations—however flawed—are my own.

The first piece in fulfillment of this commitment is coming soon. For now, a very happy Year of the Horse (and 2026) to all!1

About Here It Comes

Here it Comes is written by me, Graham Webster, a lecturer and research scholar at the Stanford Program on Geopolitics, Technology, and Governance, and editor-in-chief of the DigiChina Project. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

I’ve gathered some insights over the years, in significant part from discussions in academic or think tank settings that happened under Chatham House Rule or other restrictive ground rules. I wanted to understand more of the story and share it, because I believe that while it was widely reported and discussed, people often don’t know why this happened and certainly don’t get much of how. It turns out to be a pretty dramatic story—a small group of policy entrepreneurs with a then-rare focus on national security and AI join with a US-China tech thicket already under way and get something big done at what counts as light speed for Washington.

Today, after interviews with more than 10 former officials and experts, WIRED published my attempt to get the basics of the story out there. Some sources I had heard from in private settings, and I asked if they’d speak with me for this story. Others I reached out to for the first time, and they agreed. I’m grateful for all the time and insights they shared and expect more work to come out of this.

For now, head over to WIRED and check out the story.

Final note: Before I was whatever combination of scholar and wonk and writer I am now, I was a journalism major. And before that I was a tech enthusiast. It was a blast and a dream to work with WIRED on this, and I’m grateful to all the editors who had a hand in the piece, especially Louise Matsakis, who took the chance.

Now I hope you’ll all read and tell me what I’ve got right and (hopefully minimally) wrong, and what else you’d like to understand about this.

Subscribe now

Share

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

For the Tech Policy Press podcast, I spoke with Editor Justin Hendrix. Listen and read the transcript at their site:

https://www.techpolicy.press/unpacking-chinas-global-ai-governance-plan/

Here’s TPP’s intro:

On Saturday, July 26, three days after the Trump administration published its AI action plan, China’s foreign ministry released that country’s action plan for global AI governance. As the US pursues “global dominance,” China is communicating a different posture. What should we know about China’s plan, and how does it contrast with the US plan? What's at stake in the competition between the two superpowers?

To answer these questions, I reached out to a close observer of China's tech policy. Graham Webster is a lecturer and research scholar at Stanford University in the Program on Geopolitics, Technology, and Governance, and he is the Editor-in-Chief of the DigiChina Project, a "collaborative effort to analyze and understand Chinese technology policy developments through direct engagement with primary sources, providing analysis, context, translation, and expert opinion." Webster attended the World Artificial Intelligence Conference in Shanghai. I spoke to him right after his return.

If you’re interested in timely and often critical views of the tech and AI policy landscape, you’ll certainly want to subscribe to their podcast. Meanwhile, let me know what you think about this one!

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

“China thought it had a truce with the US. Then Trump dropped two bombshells,” says a CNN headline writer. “Export controls are endangering the fragile US-China truce,” says NYT. says “US-China Geneva ‘détente’ may be breaking” and then rightly comments that “the ‘détente’ is not really a détente and, whatever it is, it is very fragile.”

Subscribe now

The Geneva deal, at the basic level, was not a “truce” but a unilateral US retreat met with a removal of Chinese retaliatory measures. Trump backed down on the tariff front in specific, and the Chinese government pledged to remove its tit-for-tat measures responding to unilateral US moves. That’s it. Neither side got anything else except an agreement to talk more.

Other policies dating to the first Trump administration and the Biden administration that put US-China economic ties, especially in tech, on a perilous footing remain untouched. Among them:

—Chip controls. The United States maintains and continues to tighten export controls that seek to hold back Chinese development in AI among other endeavors. Sources that suggested the Trump administration might not continue tightening chip controls for AI turned out to be wrong. That was before Geneva. The joint statement following Geneva said nothing about rolling back or pausing this kind of thing, and it should be no surprise that the US government reportedly took another step on semiconductor design software this week. Bottom line: The ratchet forward of US efforts to stymie China’s advanced computing sector, a major irritant to China that even US observers have called a “declaration of economic war,” continues apace.

—Critical minerals. It was never clear what if any restrictions on the export of critical minerals or rare earths to the United States China would reverse or pause as part of the Geneva deal. The joint statement said China would “suspend or remove the non-tariff countermeasures taken against the United States since April 2, 2025.” This might include the April 4 export restrictions on rare earths, but those measures could just as well be understood to be responses not to US tariffs but to the older, ongoing, and not-suspended chip controls. They don’t seem to have been removed, and it was never clear this was part of the deal. Meanwhile, China’s 2024 ban on gallium, germanium, and antimony exports to the United States is obviously out of scope and still intact. Bottom line: China isn’t giving up on weaponizing a bit of interdependence back at the United States as the Trump administration continues down Biden’s road on tech controls. US strategists won’t and shouldn’t stop worrying about mineral supplies.

Then you have the general, continuing US-based shocks to bilateral ties. When Secretary of State Marco Rubio announced that the United States would start “aggressively” revoking visas for Chinese students, he escalated Chinese citizens’ personal safety and everyday life concerns. AP reports that in the ‘23–24 school year, 270,000 Chinese students were studying in the United States. That’s a lot of people back home one- or two-degrees separated from someone who at minimum fears their future plans could be upturned at any time.

They can take no comfort from zero-credibility modifiers in Rubio’s announcement, noting they’d target Communist Party links or “critical fields.” Capricious, sloppy, illegal operations against non-citizens are a hallmark of the Trump administration, and Rubio personally owns the abuse of power to target individuals based on false characterizations and protected speech. It is fully rational for every Chinese student in the United States now to fear being taken off the street by masked “homeland security” agents at random.

This is a new US escalation that builds momentum toward full decoupling and comprehensive rivalry, with no braking mechanism in sight. But it has nothing to do with the unilateral tariff retreat the US made in Geneva.

There was no US-China truce. If markets and pundits can’t see past panicked Trump administration efforts to obscure their direction of travel and keep it off balance sheets, that’s their problem.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

On May 1 I wrote that “it feels a bit much to say China’s government ‘appears ready for trade talks.’” I was objecting to reports that read between the lines of two Chinese government–linked social media posts and identified an apparent opening. So it is fair to say, after the Switzerland meetings between US and Chinese officials and today’s joint statement (en, zh) on a partially-temporary reduction in tariffs: Didn’t see that coming!

What happened

The concrete outcomes of the Geneva meetings are limited but significant:

• The two governments each committed to lower tariffs and then establish “a mechanism to continue discussions (协商) about economic and trade relations” led by Vice Premier He Lifeng on the Chinese side and Treasury Secretary Scott Bessent and US Trade Representative Jamieson Greer on the US side. There is no timetable for this mechanism, and it is explicitly noted that working level contacts may also continue.

• The US commitment to lower tariffs comes in two parts:

  • A 24% decrease that is to be effective for 90 days.

  • A 91% decrease that, according to the plain language of the joint statement, is permanent. This is composed of removing a 50% hike from April 8 and a 41% hike from April 9 during the tit-for-tat period last month.

  • (A 20% tariff Trump imposed citing the fentanyl issue, and 10% of the April tariffs remain, for a total of 30%.)

• The Chinese commitment to lower tariffs is parallel (indeed, unlike the original US action, it is reciprocal in the conventional, non-Trumpian meaning of the word):

  • A 24% decrease that is to be effective for 90 days.

  • A permanent elimination of a 50% hike from April 9 and a 41% hike from April 11.

  • China additionally agrees to remove or suspend unenumerated non-tariff countermeasures since April 2.

  • (A 10% reciprocal tariff remains.)

So the two sides cut overall tariffs by 91% in both directions and, for a period of 90 days, cut them 24% further. That 24% appears to be ready to “snap back” in mid-August.

Some things I seem to have gotten wrong, and right

(1) Attitude problem. I argued that China’s message was that negotiations would be possible when the US side showed a different attitude. I didn’t think that was happening, and, though I didn’t put it in my post, I did’t think it was about to happen. I certainly didn’t see it coming that Trump would publicly state that (lower but still very high) 80% tariffs “seem[] right,” but in many ways this public back-off before the high-level meetings may have fulfilled a Chinese desire to see concessions before any talks. After the Trump 80% statement, it seemed a lot more likely talks were coming. But I also failed to anticipated that the US administration would back off as far as it did.

(2) Word games about “talks,” “negotiations,” “dialogue.” In discussing Chinese denials of US statements that talks were under way, I wrote, “My hunch has been that there are indeed talks between US and Chinese government officials, but from the Chinese perspective these are working-level conversations about the logistics and conditions of potential future negotiations.” This seems to have been right.

The Ministry of Foreign Affairs spokesperson and others repeatedly made variations of a statement that “[d]ialogue and negotiations (对话谈判) must be equal, respectful, and mutually beneficial.” Neither the joint statement nor China’s Commerce Ministry statement refer to the Geneva meeting as “negotiations,” but the Commerce Ministry does characterize them as “equal dialogue and consultations” (平等对话协商). Here I think I was basically right, though I hadn’t foreseen the US side making concessions that would make the Chinese side call their dialogue “equal.”

(3) Chinese side less petty than expected. I wrote that US public statements that talks were ongoing while, probably, they were being negotiated would likely “piss off” Chinese diplomats. I’m pretty sure I was right about that! My embedded assumption, however, was that this would be a barrier to getting something like the Geneva meeting, and there I was wrong.

What have I learned?

In addition to a reminder that I don’t like predicting things and might do well to be more careful in that regard, I have learned:

• When goes all caps in a headline, he might be right, even if it seems flimsy to me based on the available text!

• Don’t take prideful public statements by the Chinese government OR Trump’s US administration at face value. (Duh, though I apparently need to be reminded that some texts do not gain value with close reading!)

• The economic pain and political consequences of total US-China economic decoupling appear to be too much for either government to bear, even if backing away from the brink makes them look weaker. (This is especially true for the Trump team, who retreated from rash moves with ZERO public concessions from the Chinese side. China only reversed retaliatory measures and changed nothing else.) Trump et al. appear to have fought to get the Chinese side to the table.

• The Chinese government never said they wouldn’t talk, only that there had to be vague appropriate conditions. The Trump team appears to have been willing to walk through that door.

• Speculation about bilateral conversations where I have no inside information is messy.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

I could be wrong, so to quickly present some of the cases people have made:

• , a former Xinhua staffer who then went on to the Beijing think tank Center for China and Globalization and is now a Princeton graduate student, writes in his newsletter that “CHINA APPEARS READY FOR TRADE TALKS WITH U.S.”
  
  He cites the Yuyuan Tantian post on Weibo, the contents of which I’ll discuss below. He also cites the “Chairman Rabbit” social media account, which is less/not authoritative but more directly supportive of talks.

• A trio of good reporters from the FT report the Yuyuan Tantian post as a “gesture” and lead with “Chinese state media has said there would be ‘no harm’ in holding trade talks with the Trump administration, indicating a softening of Beijing’s position as both sides look for a way out of their crushing tariff war.” They quote Wang, as well as Trivium China’s Andrew Polk, who I take quite seriously as an analyst.
  
  “By painting the US as the more eager, more anxious, more pressured party they are trying to portray themselves as coming from a place of strength,” Polk tells FT. “This should play well with the domestic audience and give them cover to start negotiations.”

The quotes both of these (among others) picked out from Yuyuan Tantian do seem to support the idea that there’s a Chinese government olive branch here. Here’s the original and FT’s translations:

• “但是美国如果希望同中国接触，现阶段对中国来说也没有坏处”
  “But if the US wishes to engage with China, there’s no harm in it for China at this stage”

• “谈，大门敞开，打，奉陪到底。”
  “If it is talks, the door is wide open. If it is a fight, we’ll see it through to the end.”

Subscribe now

Context gets in the way, for me

So why do I feel a bit hesitant? When I opened the underlying piece from Yuyuan Tantian, I found a relatively haughty piece that spends most of its time talking about how screwed the Americans are after Trump’s gambit, how the markets have been displeased, and how the administration has had trouble reaching deals even with more friendly governments.

The attitude is, as Polk suggests, one where the Trump people keep pathetically reaching out to talk, but honestly there’s not much to talk about unless they can get their head straight and clean up the mess they’ve unilaterally made.

In that context, it feels a bit much to say China’s government “appears ready for trade talks” based on one post by a popular audience-focused voice at the margins of authoritative state media and another by a prominent individual whose voice only carries any weight as an indication of government views because, in being allowed to be a prominent commentator on sensitive issues, they logically must have some coordination with propaganda/censorship authorities.

I got even more skeptical when I went to compare this language with the government’s prior position. Assuming today’s messages are a government signal, the relevant contrast is with official government denials that talks were under way, even as Trump insisted they were. On March 24, we have this excerpt from the Ministry of Foreign Affairs briefing (my translation):

凤凰卫视记者：近来美方不断有消息称，中美之间正在谈判，甚至将达成协议。你能否证实双方有没有开始谈判？

Phoenix TV journalist: The US side recently keeps saying China and the US are in negotiations, and even that they will reach an agreement. Can you confirm whether the two sides have started negotiations?

郭嘉昆：这些都是假消息。据我了解，中美双方并没有就关税问题进行磋商或谈判，更谈不上达成协议。

Guo Jiakun [MFA Spokesperson]: This is all fake news. According to my understanding, China and the US have not engaged in consultations or negotiations on the tariff issue, let alone have they reached an agreement

这场关税战是美方发起的，中方的态度是一贯的也是明确的：***打，奉陪到底；谈，大门敞开。*** 对话谈判必须是平等、尊重、互惠的。

This tariff war was started by the US side, and the Chinese side’s position is consistent and clear: ***If it’s a fight, we’ll see it through to the end. If it’s talks, the door is wide open.*** Dialogue and negotiations must be equal, respectful, and mutually beneficial.

Does the quote I bolded above look familiar? That’s because it’s the very same language people are interpreting as an olive branch—except it originally appeared in China’s most vociferous denial of US suggestions that talks were under way. (I reused FT’s translation to emphasize that it’s the same.)

That wasn’t the last time MFA denied US statements that contacts were ongoing. On March 28, after Trump told Time Xi had called him, the spokesperson said, “According to my understanding, the two leaders have not spoken recently.”

What’s going on here?

My hunch has been that there are indeed talks between US and Chinese government officials, but from the Chinese perspective these are working-level conversations about the logistics and conditions of potential future negotiations. The MFA language does not deny that there’s any contact between the US and Chinese governments. “Negotiations” or 谈判 are likely regarded as an official meeting on a particular matter, rather than some diplomat’s contact with some staffer on the conditions for such a meeting.

It would be utterly consistent with the Trump administration’s general dishonesty to either fabricate talks completely or embellish working level logistics calls where the Chinese side says “no talks until you get your head right” into “talks.” But that kind of public embellishment would predictably piss off Chinese diplomats, who would resort to language like “fake news” and strong denials to counter it.

Nothing about the social media posts reported today substantially changes my hunch here, though—as always—I could be wrong. That said, my basic interpretation is that the governments very likely have been talking about how to talk for weeks—just that one of them doesn’t want to talk about it.

Subscribe now

Share

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

On Bluesky, NPR’s Bobby Allyn confirmed: “all sides agreed on a TikTok deal, but after tariffs, China indicated to ByteDance it would not be signing off on the agreement. I’m told China will be holding deal hostage until they can extract some kind of concessions.”

Allyn has more: “The deal was structured to allow BD to own and maintain TikTok's algorithm. The new U.S. entity was set to license the algorithm from BD.,” Allyn wrote. “Was thought this would circumvent export-control laws, which are open to Beijing's interpretation. In wake of tariffs, China changed its stance on this.”

A few thoughts.

1. This makes it clear that, despite Trump styling himself as the dealmaker, “all sides” includes the Chinese government. No path forward is unobstructed without their consent.

2. It’s important to maintain a healthy skepticism about how close a deal really was, if Chinese government sign-off was not locked. Anyone who was saying it was done or nearly done—their versions of events need to be properly discounted.

3. If dealmakers thought they were going to avoid Chinese government scrutiny by licensing, rather than purchasing or transferring, the algorithm, they are not as clever as they thought they were. Much attention has been paid to the inclusion of recommendation algorithms on a catalogue of items that may be subject to export controls.1 But this specific legal mechanism is not the crucial point. China’s government has both legal and extralegal tools at its disposal to stop any deal if it wishes.

   1. There are probably even more direct legal levers I’m not thinking of, but another one comes to mind. The Data Security Law in Article 26 states that “the PRC may take reciprocal measures” against any country that “adopts discriminatory prohibitions, restrictions, or other similar measures against the PRC relevant to investment, trade, etc., in data, data development and use technology, etc.” What is a discriminatory prohibition, restriction, etc.? That’s up to the state. What kind of reciprocal measures might be taken? Unspecified. This is a provision sitting there to hit back when China’s government judges that it has been hit and using this provision would be convenient. There are others.

   2. Meanwhile, legal or not, ByteDance knows it can’t just disregard the state’s preferences if it wants to keep making gobs of money. See what happened to AliPay when Jack Ma resisted the state’s financial regulatory vision, or to DiDi when it went forward with an IPO after being told not to. Chinese firms are not all arms of the party-state, contrary to popular DC talking points, but they are subject to party-state’s exercise of power, legally or otherwise.

4. Meanwhile, back in the swamp, licensing an algorithm maintained by ByteDance, as recently noted, appears to be a flagrant violation of the TikTok divest-or-ban law. The “qualified divestiture” that the president must determine has happened to avoid the ban “precludes the establishment or maintenance of any operational relationship between the United States operations of the relevant foreign adversary controlled application … including any cooperation with respect to the operation of a content recommendation algorithm.” Pretty darn clear.

   1. In this whole discussion, the definition of an algorithm, or a “content generation algorithm” or a “personalized information push service technology based on data analysis” is vague. The specific architecture of the MacGuffin in this drama really does matter. I think it’s unlikely, but if the “algorithm” were frozen in time and licensed, never to be updated by ByteDance but never getting “exported,” maybe they would have found their hack. (It’s beyond my knowledge and the scope of this post to assess whether licensing would count as an export, but I do welcome experts to weigh in!) Anyway, Chinese government leverage discussed in item 3 above would still be present.

   2. It’s a US legal expert question what kind of risk Apple and Google and TikTok’s infrastructure vendors would be exposed to if they move forward in violation of the letter of the law but with explicit presidential blessing. But it still seems pretty risky to me, given the capriciousness of this president and the fact that, generally, we do change presidents in the United States.

5. If somehow a deal like the one being described goes through, US government oversight on the algorithmic manipulation and data security risks motivating the law will likely be less than would have occurred under Project Texas. We never got full details, but the proposal as reported by people who were briefed by TikTok included a bunch of government auditing and supervisory structures.

Whatever happens with TikTok, whether a deal goes through contrary to the law, or even if a ban is enacted or a truly “qualified divestiture” takes place, this whole saga will at best have only addressed the national security, privacy, and algorithmic manipulations of one app. Threats to US users’ privacy, bad behavior in the market, national security where it is implicated—the whole range of TikTok concerns and more—remain broadly unaddressed if the app at issue is not TikTok. Thats why, for years, I and others have advocated a regulatory approach that covers all services.

That’s all for now. Signing off for the weekend.

Subscribe now

Share

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Subscribe now

They frame the essay in a way that pushes my buttons, arguing this “may mark the end of the dream of a global Internet, in which everyone shares the same services.” The Internet, of course, is a constellation of computer systems connected by specific protocols—not the dominance of specific companies. As they acknowledge, there can still be a global Internet without satisfying a CEO’s dream of global revenue and market share. This pet peeve of mine aside, they describe well a shift among CEOs from portraying their dream in terms of the liberating power of technology to embracing geopolitical competition or nationalism.

Reading the essay made me think of another dream under threat.

The dream of a coalition of techno-democracies

Mainly during the Biden era, strategists advocated for a coalition of techno-democracies to counterbalance China and a potential bloc of techno-autocracies. At Foreign Affairs in 2021, Justin Sherman and I wrote skeptically of this idea and in support of a techno-globalism we already believed was under threat: “Embracing a trend toward politically delineated technological ecosystems,” we wrote, “will undermine the open ethos that fuels and benefits freer societies—and bolster the top-down, controlling ethos favored by repressive regimes.” We rather idealistically (dreamily?) argued that the global reach of US firms led to a global responsibility: “Cyber-utopians once dreamed of liberation spreading from an Ethernet cable; now Washington must ensure that its companies don’t spread exploitation and insecurity instead.” In our view, the US government needed to rein in its own tech firms in alignment with democracy and human rights to avoid bad outcomes both at home and abroad. The context would be inescapably global.

The Biden administration pushed hard to foster democratic solidarity, including leading “The Declaration for the Future of the Internet,” which had global ambitions but was announced with 60+ exclusively democratic partners. The declaration contained some laudable principles, but they were never seriously applied to discipline US firms in the way Justin and I would have hoped. The Biden administration never resolved a tension between reining in big tech and unleashing it to beat China.

Now the world is also trending away from the vision of a techno-democratic coalition, which although never achieving its highest ideals was not without strategic importance. As the US government turned decisively to “weaponize interdependence” (in Farrell and Newman’s memorable coinage) it became more important to the United States to bring some allies along, and it became more likely China would further its own weaponization of supply chains, which was once mostly a matter of what-ifs. Some key allies were crucial in reaching what may end up marking the high point of this techno-democratic alignment, when the Biden administration successfully put the screws to Japanese and Dutch officials to deny China access to semiconductor manufacturing equipment after unilateral US controls introduced in 2022. US strategists believed it was so important to hold China back in advanced computing and artificial intelligence that they pushed through ground-breaking export controls. Nearly everyone believed the effort would be self-defeating without others’ help, but the United States got that help—for now.

Going forward, it is reasonable to expect that China’s government will leverage its position in ways that will require even more coordination to withstand. So the big question is not just whether Meta and Alphabet get hosed in the European market—potentially a rather small economic or political matter compared to the effects of tariffs announced this week. The question is whether countries that, due to their regime type, are putatively natural US allies will be less likely to lend a hand in the future. Thinking pragmatically, could an unreliable United States have created for itself a more determined Chinese competitor, one that can make more credible assurances to erstwhile US allies? Thinking from a stance of democratic solidarity, will governments view a US government that increasingly sides with Russia against Ukraine and thrashes rule of law as a member of the democratic club, an enabler of autocratic power, or something else entirely?

No one can reliably predict macro geopolitical events at this stage, so the take-away has to be uncertainty. But the uncertainty can be a reminder of frames to question when they come back around. When policy advocates talk about democracy vs. autocracy, it cannot be assumed where the United States as a country or the political regime controlling its government stands. When people call for the development of “democratic AI,” it cannot be assumed that systems from the United States or controlled by US actors are in any meaningful sense democratic. If a self-styled coalition of “democracies” includes the United States, it behooves us to ask what actually is the commonality unifying the group.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Subscribe now

Former US President Joe Biden left office with an urgent warning about the power of money and the technology industry. In his farewell speech, Biden said that “an oligarchy is taking shape in America” and compared the concentration of wealth and power to the 20th-century robber barons. It was a general warning with broad resonance, but Biden was not done, and he turned specifically to the dangers of the tech industry.

He invoked perhaps the most famous presidential farewell since Washington—Eisenhower’s 1961 warning of an emerging “military-industrial complex” and the “potential for the disastrous rise of misplaced power.”

“Six decades later,” Biden said, “I’m equally concerned about the potential rise of a tech-industrial complex that could pose real dangers for our country as well.” He expressed concerns about misinformation and disinformation, alluded (without naming names) to Meta’s announcement that it was ending fact-checking in the US, and called for “hold[ing] the social platforms accountable.”

In the next breath, however, Biden turned into a nervous cheerleader for those very platforms, at least the ones that develop AI. “Artificial intelligence even has the potential to help us answer my call to end cancer as we know it,” Biden said. “But unless safeguards are in place, AI could spawn new threats to our rights, our way of life, to our privacy, how we work, and how we protect our nation. We must make sure AI is safe and trustworthy and good for all humankind.”

Biden’s hedging, citing both great potential and grave risks, finally gave way to a clear statement in support of US AI development: “In the age of AI, it’s more important than ever that the people must govern. And as the land of liberty, America—not China—must lead the world on the development of AI.”

In just a 10-paragraph passage of his speech, Biden brought forward two visions of technology that both flowed through his presidency. On one hand, reining in the power of tech platforms and beginning to erect safeguards around AI development was a major theme. In this vision, technology and capital run amok could harm individuals, markets, and security. On the other hand, the Biden administration took extraordinary measures through its last days to ensure US technology and the associated centers of capital are ahead of their Chinese counterparts. In this vision, especially in AI fields, the US must empower tech companies and encumber Chinese competitors to ensure democracy, prosperity, and national security.

These ideas are not entirely incompatible. One can favor certain limits on tech platforms while seeking to enable their innovations through light regulation or direct support. In their application, however, these visions conflict when it comes to how to treat companies and the wealthy individuals connected to them.

A vision identifying unchecked social media platforms as threats to democracy and abusers of market power animated the Federal Trade Commission under Biden-appointed Chair Lina Khan and antitrust enforcers in the Department of Justice. The business model that profits from gathering intrusive data on users and leveraging it for ad targeting or outright sale has never been under such frontal attack in the US. Corporate and investor interests were so unhappy with this thrust from the administration that some reportedly lobbied Vice President Kamala Harris to oust Khan if she became president. It was a kind of tech populism wrapped in hotshot lawyering, and it drew enthusiastic support from digital rights supporters and figures on the left. Even some on the right broke with pro-corporate GOP norms to support Khan’s aggressive approach.

Meanwhile, the Biden team put a great deal of energy into another of its tech visions, the need to build and maintain as wide a lead as possible over China in “force multiplier” sectors, including semiconductors, quantum, AI, biotech, and clean energy. When National Security Adviser Jake Sullivan presented this case in 2022, he did so at a conference organized by the Special Competitive Studies Project (SCSP), a subsidiary of the Eric & Wendy Schmidt Fund for Strategic Innovation. Eric Schmidt, of course, is a former CEO of Google and an important policy entrepreneur advocating for greater US government attention to the national security implications of AI and tech competition with China, among numerous other projects. He is also an active investor in AI companies. Meta CEO Mark Zuckerberg, who must be understood as one of Biden’s targets in criticizing the social platforms, also has for years cited competition with China as a reason not to, in his view, over-regulate US platforms. Meta reportedly lobbied quietly but doggedly to ban its Chinese-owned competitor TikTok.

These billionaires are not the only advocates for a strong competitive stance toward China, and their policy stances are not unique to those with vested interests. Concern about Chinese technological power is widespread, especially among people who believe AI systems may soon vastly outstrip human capabilities and revolutionize military power. China’s offensive cyber operations, disclosed by the Biden administration, underline the degree of tech-related peril the US faces in the event of large-scale conflict, and Sullivan’s “force multiplier” frame for key technologies is not without merit. At the same time, there is an unavoidable tension between warning of a tech-industrial complex and advancing policies that are developed with the support of tech moguls and stand to benefit their bottom lines.

The administration’s approach to AI is another example of this tension. Biden’s centerpiece action was the executive order on “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,” which incorporated thinking from critics who have long identified abuses of machine learning and other automated systems, as well as concerns that preoccupy many of the technology’s boosters, namely that future AI systems could bring about catastrophic risks. The order sought to ready the government for an AI era, and substantial work went into developing safety testing for AI models. Yet the administration did not take or even strongly advocate binding action to rein in discrimination, bias, or violations of privacy or intellectual property rights. Top executives from developers of large AI models advised the administration and participated in voluntary commitments, but they remained free from substantial binding regulations or safeguards.

The concrete outcomes from these competing visions are broadly in question as the new administration unfolds. President Trump immediately rescinded the AI executive order, Khan is out as FTC chair, and the fate of the Biden administration’s export controls targeting China’s semiconductor industry is uncertain. There was no apparent concern about a tech-industrial complex on Inauguration Day in the Capitol Rotunda, where multiple tech billionaires were seated in front of Trump’s Cabinet nominees.

Like the World War II General Eisenhower, the elder statesman Biden has seen the tensions behind his warning first hand. With luck, Biden’s warning will be more successful than Eisenhower’s.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Reminder: Thanks for reading. Please don’t hesitate to share reactions or objections with me directly or on Bluesky, and please consider sharing with interested friends and colleagues. If you haven’t subscribed, sign up here for my ongoing work to understand the interplay of US-China relations, technology, and climate. If you especially wish to support this work and have the means, please consider a paid subscription.

Subscribe now

Share

1. The Court’s choice to ignore the divest-or-ban proponents’ concerns about information manipulation may be legally sound, but it isn’t good history.

Standard disclaimer: I am not a lawyer or a freedom of expression scholar, and although I have spent a great deal of time surrounded by lawyers and law professors, I’m an advanced amateur and welcome correction. (I think I have this bit right, though.)

In order to judge whether the First Amendment should block the TikTok divest-or-ban bill, the Court needed to decide whether to apply “strict scrutiny” or “intermediate scrutiny” to the law’s provisions. Under strict scrutiny, among other things, precedent would call for the government’s goals to be accomplished with the “least restrictive” means in terms of limits on speech in order to be constitutional. Under intermediate scrutiny, the law merely need not “burden substantially more speech than is necessary” to advance those goals.1

The lower court assumed strict scrutiny applied and decided the law passed muster. SCOTUS instead applied intermediate scrutiny. In doing so it said “the challenged provisions are facially content neutral and are justified by a content-neutral rationale” (10). This matters because applying intermediate scrutiny frees the Court from evaluating whether the law is the least speech-restrictive option for achieving government goals. Concretely, this means the justices can ignore the question of whether a law mandating something like Project Texas—which would have installed controls on user data and algorithm management at TikTok US ultimately overseen by the US government—would have addressed the goal of protecting national security.

The Court makes a pretty slick move to apply only the lower standard. First, it reasonably argues that one of the two driving concerns behind the law is content neutral: protecting US user data from possible Chinese government access. Second, it dodges the question of whether the other main concern is content neutral: the risk that China’s government might manipulate feeds for propaganda or information interference goals by pressuring TikTok’s Chinese parent company or its employees. The Court simply declares the second justification is irrelevant, ignoring its much more content-related logic—therefore avoiding applying strict scrutiny.

To eject the feed manipulation concern from their analysis, the justices assert that “Congress would have passed the challenged provisions based on the data collection justification alone” (18) On Bluesky, University of Chicago Law School Professor Genevieve Lakier wrote that this “gives govts space to whitewash bad content-based motivations by tacking on plausible-sounding content-neutral ones.”

I’ll leave that to legal experts, but I do have something to say about the argument itself that the law would have passed and been signed without the further propaganda concern. As someone who has watched the TikTok ban story closely for nearly five years, this does not sound like a strong argument. Indeed, while both data protection and algorithmic manipulation fears have been present all along, the data concern was dominant in the discourse at first. That argument met significant friction, however. Many experts have made the cogent argument (which I have also advanced) that banning TikTok wouldn’t at best only partially address the real risk that China’s government might gain access to large-scale US personal data and try to leverage it. Addressing this risk calls for more comprehensive data security legislation that would curb risks of personal information transfer through the data broker economy or through badly secured platforms Chinese security services could simply hack into. A comprehensive policy would certainly have something to say about TikTok’s structure and data access provisions, but it would focus on the data exfiltration problem holistically rather than zeroing in on one of many sources of risk.

As more and more people understood the data problem was too broad to justify a unique focus on TikTok, proponents of the bill before it was passed last spring shifted their attention to the argument that in an emergency, for instance a Chinese invasion of Taiwan, Bytedance might tweak the algorithm to deliver content to millions of US users supporting China’s position. The risk of covert influence was the big red flashing light in the discourse as the bill passed. Data security was still there, but it’s just not reasonable to exclude the other, more plausibly “content-based,” justification from a history of what got this long-running effort over the finish line. Maybe the Court has blinders here due to the record, but its argument is not valid outside their narrow context. Both justifications motivated Congress and Biden. (Whether that means SCOTUS should have applied strict scrutiny I leave to First Amendment scholars.)

2. The United States doesn’t get to claim that it’s against national security–driven Internet censorship anymore.

It’s not clear today whether TikTok will indeed be banned, or if so for how long, but the three branches of the US government have spoken: Speculative national security concerns based on nationality of a company’s ownership are legitimate grounds to erect formidable barriers preventing US citizens from accessing information on or expressing themselves through that company’s product. In other words, Congress, the president, and the Supreme Court endorse censoring websites or apps on the Internet inside the United States if they say there’s a national security risk.

To be fair, there are several arguments against the way I have just framed this:

• First, one could argue the Supreme Court was at pains not to endorse censoring anything but TikTok. This is true. They write: “A law targeting any other speaker [i.e., other than TikTok] would by necessity entail a distinct inquiry and separate considerations” (12–13). Yet the law at issue gives the president the ability to designate other entities for the divest-or-ban treatment, and today’s decision signals broad deference to the government’s assessment of national security risks. There’s no obvious barrier to growing the censored list.

• Second, one could argue the result is not censorship, because there was the option of divesting. We could debate whether divestiture was ever a real option, but at present there is no concrete sign that it will happen. (Who knows, though!) At least for the moment, the effective outcome of the law is massive-scale censorship in terms of US citizens’ access to information and their choice of venue for speech.

• Third, one could argue that the national security concerns were so intense and specific, including necessarily undisclosed classified evidence, that this is a narrow action to address a serious threat and doesn’t signal a new comfort with Internet censorship. There are multiple weaknesses to this objection. At the Supreme Court level, the justices explicitly state that they relied only on the public record and not on the classified evidence (13), and the public record is pretty speculative. Meanwhile, the national security dangers identified are broader than TikTok, and it’s hard to take this seriously as an effort to address them absent a corresponding comprehensive effort. Finally, to deny that a TikTok ban is censorship is to deny that cutting off access to a huge range of information for one-third or more Americans is censorship, and I just don’t accept that. You can argue that it’s justified, but the interference with freedom of expression and access to information is undeniable.

So I believe my framing stands, and proponents of the law, the rest of us in the United States, and people around the world need to reckon with a massive shift on the part of the US government regarding Internet censorship. The US government has long been a loud critic of efforts like China’s Great Firewall, which maintains an online border to cut off citizens from large amounts of information in the name of public safety, national security, and sovereignty. Now, the US government is on the verge of erecting the first span of what could become our counterpart. (Call it a Cyber Border Wall, perhaps—a national security–motivated effort, still porous, but aimed at keeping out a reputed foreign menace while sacrificing erstwhile US ideals of openness.) The TikTok law is one tool the government might use. The ICTS rule developed over the Trump I and Biden administrations is another.

As I said this morning in resharing a post by David Kaye, a UC Irvine law professor and former UN special rapporteur on freedom of opinion and expression, this watershed moment in the US system’s willingness to censor the Internet is the biggest take-away for me today. Kaye said it better than I could: “can the United States ever again, with a straight face, argue that another country's internet shutdown, website blocking, app-banning, etc is a violation of the global right to freedom of expression? that it's unnecessary for national security? no. imo its prior moral authority is getting shredded.”

—

That’s enough for now. Like it or not, there will be more to discuss. And whether one likes or dislikes today’s outcome, it’s certainly a big deal.

Subscribe now

Share

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Subscribe now

The case for a TikTok ban has been multifaceted, but for the most part it has relied on imagined potential future harms. Today, writing a thread on what’s next for the ban, I found myself imagining some scary potential outcomes after a ban takes effect.

First, to recap, many ban proponents imagine:

• The Chinese government might exfiltrate data at scale and use it somehow to harm US national security. Location data, browsing data, social graph, misusing access to microphones or cameras on devices—all kinds of possible concerns. The fact that such large-scale misuse has not been documented isn’t relevant here, because it’s about what China might do. (And the fact that the company has done shady company things like try to track journalists’ locations in a leak investigation, or leave open access to Beijing engineers, doesn’t help their case.)

• The Chinese government could use physical access to ByteDance systems or personal leverage over staff to try to manipulate US public opinion. Evidence of this kind of thing being under way is quite thin, but some bleed between China-based content moderation practices and the global app has been reported, and the company’s assurances that mistakes were made and fixed aren’t convincing if you already don’t trust them. Moreover, you don’t need evidence of such a campaign now if you’re imagining this would happen in the context of a future contingency—the classic example being a Taiwan contingency.

• Persistent integration between Chinese and global software teams could give Chinese government actors a chance to insert or exploit cyber vulnerabilities.

I have long argued that these risks are indeed imaginable and worth taking seriously, but that they are largely not unique to TikTok and ought to be dealt with on a comprehensive basis covering all such services. Where there are risks specific to TikTok, I have argued that the Project Texas mitigation proposals, while not publicly detailed in full, looked like they might do a good job on all of this. I’ve never found it useful when people try to refute the possibilities, because you can’t prove wrong a speculated future. Anyway, the law passed, and the DC Circuit says it can stand, and comprehensive legislation isn’t forthcoming. You don’t have to use your imagination about that.

Share

Now for some imagined possibilities stemming from a ban. Allowing that we don’t know yet whether a ban will actually happen (SCOTUS, Trump, a divestiture, possibly further lawsuits, all could intervene), here are some slightly screwball1 but nonetheless plausible results that could come about assuming a ban. Imagine:

• Apple and Google could comply in removing TikTok from their app stores, leading large numbers of committed users to learn how to jailbreak their mobile operating systems and side-load an internationally available copy of the app. This would mean users lose the (imperfect but important) cybersecurity refinements of an up-to-date iOS or Android version and the similarly imperfect efforts by the companies to vet apps they distribute. Users could then be compromised by cyber criminals and adversary governments. Any effort to embed hidden malware in a TikTok app would be more powerful in a jailbroken OS, which would be more vulnerable in other ways as well.

• If the ban is enforced such that DNS for TikTok’s servers is blocked or unstable, users may turn to VPNs to tunnel into countries where access is unencumbered. Yet many or most would not want to pay for their VPN, increasing the chance that the VPNs themselves would act as data exfiltrators or attack vectors.

• In anticipation of a potential Project Texas arrangement with the US government, TikTok said it would base its US user data in US-based Oracle infrastructure. Under the ban, Oracle would be prohibited from providing these services, and US user data would be stored wherever convenient for TikTok. For people who worry about where US user data is stored and how, US authorities would lose visibility.

The purpose of this is not to argue over the ban. There’s been plenty of that, and the die is cast. Instead, I wrote this up because I have been thinking a lot about the imagined potential futures that animate the policies I track on technology issues in the United States and China. What AI might be, what military contests might emerge, what might decide who “leads” in what, and why that matters… Speculating about uncertain futures is unavoidable in policymaking, but the guesses we make are a big deal, and often the imaginaries seem to be the least contested part of the discussion.

I can imagine some good outcomes of a ban. Maybe people would spend less time on a media product designed to be addictive. Maybe they’d garden and read books and go to the coffee shop or the pizza parlor to sing and dance at each other. Or not: Stock performance today suggests many believe outcomes for Meta’s products would be good.

Share

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Subscribe now

I. Climate catastrophe is still coming for us all. The United States now is likely to be even more of a spoiler of global mitigation and adaptation efforts. When climate factors add stressors to international security, markets, and populations, it is even less likely the United States and China will constructively cooperate. The world will have to try to navigate around the United States, but that might not work very well. Some recent policy improvements will be sticky, but we needed much more.

II. The specifics of US policy toward China will be very hard to predict, but in sum it will be disruptive. Deal-making and capricious targeting of punitive measures is likely. (TikTok could be spared, or not.) Overall, however, the era of selective “de-risking” is likely over and a trend toward more thorough “decoupling” is likely—even if its extent may be limited by economic realities.

III. US economic measures toward China will likely be even more disruptive than the existing tariffs, and China’s government will retaliate. How it does so is not predetermined. Will it help or hurt a company’s fortunes if its CEO is buddies with the US president? Overall, US businesses will wish to further mitigate their exposure by decreasing reliance on Chinese supply chains and markets. Trade negotiations may advance in parallel, but I wouldn’t hope for much: “Phase one” was both a joke and never worked out.

IV. The US government is likely to find it harder to bring allies and partners along on costly measures toward China. Already pushed beyond their comfort zone, how much appetite will the Netherlands and Japan have to undermine China’s chip industry “for democracy” if it’s an aspiring autocrat cajoling them? Any number of proposals in the name of democracy will be treated with deserved skepticism; democratic publics will demand this of their governments. But geopolitical bloc logic may still prevail.

V. The China Initiative, or something like it, will return and may be even more careless and damaging to innocent people. Anti-AAPI hate and violence may rise again. I hope I’m wrong.

VI. US human rights narratives about China, though often righteous on their face, will suffer even more from the perception and reality of hypocrisy. It has already been rich to hear about the rules-based order from a government unceasingly providing weapons used to kill tens of thousands of civilians. How about the religious freedom pitch from the folks who brought us the Muslim ban, or gender equality from the people denying medical care to women?

VII. Current trends in cyber operations attributed to China point to serious escalatory potential. Pre-positioning in critical infrastructure, exploiting telephone wiretap systems to target the campaigns before a presidential election, and no known bilateral mechanism for strategic stability in this area—this is volatile stuff in addition to the other present uncertainties.

VIII. US-China dialogue, official and unofficial, will be diminished compared with the hypothetical alternative future. We may still keep some momentum, and it’s hard to sink below the Covid nadir, but it will be more stressful and logistically challenging to get US and Chinese groups together to discuss problem solving or common interests.

IX. Fewer US citizens will develop up-to-date and grounded knowledge about China. Business travel, study abroad, academic fieldwork, etc., will seem more risky and less career-boosting due to a combination of repression and suspicion in China and questions about loyalties in the United States. Though still coming in large numbers, even more Chinese will think twice before studying in the United States. More US decisions about China will be made based on narratives, politics, and desk research.

X. In the United States we will hear—constantly—that problems are China’s fault, or immigrants’ fault… anyone but ours. But in most cases they will be problems created by a combination of macro realities (tech, climate, economic) and bad choices or a dysfunctional political system (industry capture, money in politics, bad information sources). Think economic inequities, job losses, housing costs, a strained care economy. Politicians in hurricane country were blaming the actual weather on Democrats.

That’s all for now. What to do about it will have to come another day. I’d love to hear your ideas.

Subscribe now

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.

Subscribe now

Since I resumed regular work travel to China last summer and during five trips since then, it has still been hard to imagine US-China futures. The overall trend lines are concerning, even as some potentially stabilizing diplomatic lines have reopened during the Biden administration after the twin lows of the Trump shock and pandemic recriminations. In the context of technology, the future appears especially uncertain after another pair of events: the October 2022 US export controls on semiconductors and chip-making tools, and the November 2022 release of ChatGPT. There was always a mixture of cooperative dynamics (e.g. in basic science and complex supply chains) and adversarial activity (e.g. in cyber operations and military tech), but the balance has been tilting toward the latter.

There’s another factor that has kept me from developing any strong sense of US-China futures, however, and it’s one I have shared with many Chinese interlocutors: The United States and China both have leadership uncertainty. It’s just that the United States knows ours is coming on a Tuesday in November, and China’s will come at some other time.

When I have shared this observation in China, generally outside of the structured “meeting” part of an interaction and during the food/beverage part, I have seen a lot of agreement. In raising the issue I am explicitly expressing my unease about what will become of the United States politically, and there’s no shortage of similar unease in China. Before the pandemic, being a person who sometimes learns by (intentionally or not) pushing at the boundaries of comfortable conversation topics, I often got the sense that talking about Xi Jinping’s un-delimited leadership was taboo—at least with a US researcher. Now, however, the obvious questions about what happens when China’s top leader is one-way-or-another no longer the boss is barely if at all below the surface.

Flying home from Beijing yesterday, two days before a Tuesday in November, I realized that uncertainty and anxiety about short- and long-term political futures is the highest I can remember in both the United States and China. We are both at our political wits’ end, strained by challenges and divides with no clear solutions. So much of the focus among observers of and participants in US-China relations has been on the mounting strain between the two countries. Yet the strain within feels even more important and pressing on both sides.

In December 2020 I wrote:

For several years it has been a truism that there is a new bipartisan U.S. consensus on China, often summarized as an agreement that competition is the dominant feature of bilateral ties. That consensus is real, as far as it goes. But competition is vague, and there is nothing even resembling a consensus on the nature of that competition, let alone on what to do about it. 

This is just as true today as it was four years ago, and what happens this week will be immensely consequential for how the US government acts upon this vague consensus. Elections and scheduled leadership transitions mean no one can credibly argue US policy on major issues will be constant long into the future. At minimum, the unknown candidates in one or two cycles could have very different ideas. As Chinese observers often understand better than Washington strategists, US commitments and goals will often be upended on a Tuesday in November.

US and Chinese analysts both have a tendency to imagine that Chinese policy orientations will follow current trendlines forever. While we all know the top leader won’t reign forever, when and how things change are mysteries that may sneak up on us. It’s hard to figure an unknown event at an unknown time into guesses about the future. Still, the consequences for how China acts in the world will be enormous.

Well, the present Tuesday in November has arrived. We’ll see what happens, and I’ll be trying to keep the dual uncertainties in mind regardless.

About Here It Comes

Here it Comes is written by me, Graham Webster, a research scholar and editor-in-chief of the DigiChina Project at the Stanford Program on Geopolitics, Technology, and Governance. It is the successor to my earlier newsletter efforts U.S.–China Week and Transpacifica. Here It Comes is an exploration of the onslaught of interactions between US-China relations, technology, and climate change. The opinions expressed here are my own, and I reserve the right to change my mind.