Nvidia and Economic Espionage

0:00Cameron Abadi: Life is all about making decisions.

0:03Adam Tooze: Sometimes you make questionable ones, like buying a new TV that's too big for the wall because the electronics store employee was a little too good at their job. Then, sometimes you make great ones, like protecting your home and auto with Progressive. You get customized coverage options, so you can be sure you have the right coverage. Get your quote today at Progressive.com.

0:25Cameron Abadi: Progressive Casualty Insurance Company affiliates and other insurers, not available in all states.

0:36Adam Tooze: Hi, and welcome to Ones and Twos, FPs Economics Podcast. Every week we take a couple data points, use them to try to explain the world. I'm Cameron Apati, FPs Deputy Editor with you in Berlin, Germany. Joining us as always is Adam Tues, FPs Economics Columnist and Columbia University Professor. Before we get to Adam, one small point. We recorded the bulk of this conversation on Wednesday. We had one data point. It is tied to the news and the news has moved on a bit since we recorded on Wednesday. Namely, the data point is related to NVIDIA. So we just wanted to give a quick update here at the top of the show before we get to the bulk of the conversation.

1:16Adam Tooze: And before we even get to the data point, one quick note. Discerning listeners might remember that we had teased a live show that was supposed to be happening this week and was supposed to be ostensibly on the feed this Friday. Of course, this is not a live show. And the reason for that is that we had to postpone it. Unforeseen circumstances intervened. But some good news is that the show has been postponed to a specific date. That is March 15th. That's more time for those of you out there who perhaps couldn't make it to this past week's show. Potentially a good Christmas present for those of you looking for one.

1:52Adam Tooze: So yeah, you can mark that date down March 15th. Apologies to those who had tickets for this week's show. We hope to see you all there. It's going to be a lot of fun. We hope to see a lot more people there as well. It's Festsaal Kreuzberg in Berlin, March 15th. The link in the show notes is live. If you want to book tickets, you can do so now. So back to that data point, it is 200 billion. That is approximately how much market value NVIDIA, that is the semiconductor company manufacturer at the heart of the AI economy around the world. That's how much money that company's stock lost just on Monday this week.

2:36Adam Tooze: That's about 4.3 % of its total value. It was, I think, at one point, the biggest company in the world. Now it's approximately the third as I record this. That was after its chief executive, Jensen Huang, disclosed a romantic relationship in a memo to employees. And the problem is that the woman with whom he was having that affair, Elise Huang, is a former executive at another company, Megaspeed, which is ostensibly based in Singapore, but some are suggesting is a cutout for Chinese technology companies. What's clear is that Megaspeed was an NVIDIA customer and that Jensen Huang seems to have been lobbying on Megaspeed's behalf with the Trump administration, which ostensibly had rules in place to prevent NVIDIA's chips from reaching China, but may have found a route through Megaspeed ultimately.

3:30Adam Tooze: A probe has been opened by the chip giant investigating whether Mr. Huang has been the target of a foreign intelligence effort. This comes as concerns grow over Huang's interactions with Megaspeed AI executive Alice Huang, an alleged Chinese intelligence asset. And so the New York Times reported that Alice Huang may have in fact been a spy for the Chinese Ministry of State Security. We can't confirm that's the case, but what's clear is that this story is not only about the AI economy or the Trump administration's susceptibility to lobbying and corruption, but moreover, a story potentially about economic espionage and the way it's intersecting with the AI economy right now and more generally how it functions in the global economy.

4:19Adam Tooze: And so we thought we'd take a step back to consider the issue from that perspective. And especially now, it's even more urgent with Jensen Huang having announced on Thursday that he's resigned, which is leading to an even greater slide in the market value of NVIDIA. So a kind of vent here in the form of this affair. And so we thought we'd dig into economic espionage and how it works. And so with that, we're going to turn to the conversation that Adam and I had on Wednesday and look forward to seeing you again next week. And of course, at that live show in March. Adam, to start off with, I thought it could be useful to kind of establish some definitions about what we're talking about here, because in reading about this subject, you come across terms like industrial espionage, economic intelligence, and those are sort of contrasted with or compared to economic espionage per se.

5:20Adam Tooze: So what are the differences between those things, industrial espionage, economic espionage, and economic intelligence, so to speak?

5:28Cameron Abadi: Yeah, I think one can distinguish usefully between the targets of data gathering, intelligence gathering, the purposes of it, the beneficiaries, and the modes of intelligence gathering being used. And industrial espionage, I think, and the definitions here are a little fluid. I mean, there's not like a canonical definition here. But I think in general understanding, industrial espionage is above all targeted at particular businesses by more often or not other businesses, but sometimes by governments. And it's about accessing commercial secrets, business secrets, business strategies, business technologies, and it involves various types of subterfuge or illegal action.

6:19Cameron Abadi: Economic espionage is sort of blurred into that zone because it also involves various types of subterfuge. But the agents more commonly involved, when we say economic espionage, one immediately thinks of governments. Why is that, I think? Because when we think of businesses, we think of private economic activity and profit. But when we use the word economy, we immediately start thinking about GDP and national interest. And so the sorts of information that are being targeted there might be more to do with technologies of general strategic interest. So that will be different from industrial espionage or purely commercial espionage.

7:01Cameron Abadi: And then economic intelligence ranges the whole gamut from, say, wartime efforts on the part of, I don't know, the British who are aiming a strategic bombing program against Germany and therefore need to understand how the German economy functions and built an economics intelligence wing of the British military war effort so as to organize that. And that ranged from famously people who were architectural historians who knew which cities would be inflammable if you dropped infantry bombs on them, all the way through to an understanding of the input-unput diagram, the supply chains. And it involves more often than not open source information.

7:40Cameron Abadi: It doesn't involve any subterfuge at all. And so economic intelligence kind of blurs into the world that we're in, of economic reporting, of newsletters, of regular updates. The Economist magazine has what it used to call the Economics Intelligence Unit, which was presumably very well-connected journalists doing studies. Now, those journalists, presumably, when they were traveling around countries like China or the wrong place of the United States, might very well find themselves being arrested on charges of economic espionage. They would have thought that was an inherent risk. Very little risk there, I would have thought, of them being accused of industrial espionage and otherwise stealing secrets for whom.

8:19Cameron Abadi: They wouldn't be stealing those on behalf of The Economist magazine. So maybe that illustrates, that spectrum kind of illustrates the distinctions here. And I think the allegation in this case, since we're hooking onto the NVIDIA story, is that it sits more squarely in the domain of economic espionage. But on the other hand, it targets a particular business and the top man in that particular business, the charismatic figurehead indeed of the entire hardware side of AI, indeed the profit of AI in general because NVIDIA's business is tied up. So it would seem that it sits somewhere on that boundary.

8:57Cameron Abadi: I don't think the allegation, however, is that Jensen Huang's lover was doing this with a view to commercially profiting herself. I don't think that's the suggestion. All of these are allegations at this point, but I think the suggestion is it's MSS, the Ministry of State Security, which is doing this on behalf of the Chinese state. So it would sit more in that second category.

9:19Adam Tooze: Yeah, I do think it's important to emphasize that we don't exactly know the facts of what happened here. You know, Huang has admitted the affair. The spy claim rests on reported intelligence assessments and China's foreign ministry has said it's all unsubstantiated. So who knows when we will exactly know what happened here. But yeah, clearly espionage of various kinds seems like unavoidable in the general economy, given that every company is looking to get an edge, but it is this kind of economic espionage per se, where governments are involved in breaking rules to get access to secrets. That seems what may be most interesting.

10:01Adam Tooze: And I'm curious, yeah, what are then some major examples of economic espionage then from history?

10:08Cameron Abadi: I mean, in the modern era, broadly understood, I think you could quite helpfully distinguish three huge long epochs. So one would be the history of mercantilism. So between, say, 1700 and the mid-19th century, where you have state-backed projects of industrial development, which are both defensive, protecting industrial development against outsiders, and proactively trying to acquire secrets so as to drive development. The era where industry per se was new, and textiles became the key center for this. And in that instance, probably the most successful and dramatic are efforts by the French and the Americans to steal secrets of modern textile manufacturer from the British, who at that point were the leading industrial power.

10:54Cameron Abadi: British, we could say a bit about this in a minute, themselves acquired these in part through encouraging migration of Huguenots, persecuted Protestant minority from France. But in any case, by the 18th century, it's the British who are the target. And both the French and the Americans, both at the level of individual entrepreneurs, but in the French case, apparently coordinated by the state, make systematic efforts to attract British expertise by hiring people, which is a tried and tested means of doing this sort of information acquisition, not as spies, but literally just hiring them away. So they bring their trade secrets with them.

11:29Cameron Abadi: But in some cases, also something that's much closer to spying, where folks would visit British factories, be taken around, apprentice themselves, and then take those secrets home with them. And in fact, would they dedicate themselves in the case of Samuel Slater, and indeed Lowell, Francis Cabot Lowell in the American case, these men would actually train themselves so as to be able to go into a factory, see a machine, leave innocently with no pen and paper or anything like that, but go home to their, you know, whatever inn they were staying at and quickly scribble down what they'd just seen, and then take this information out of Britain, making sure that their baggage wasn't searched.

12:07Cameron Abadi: So this was quite serious stuff. And amazingly, economic historians have actually begun to investigate what economic impact this had, because these stories are fun and glamorous, but do they actually matter? And a recent economic paper with perhaps my favorite economic history article title ever, which is You Only Weave Twice, Espionage and Industrialization in 18th Century France. Obviously, a play on the famous James Bond film by an academic called Bruno Caputini, University of St. Gallen, actually traces the impact, because this is the sort of thing that economic historians and economists love to do, what is the shock effect of the treatment of a French textiles community by the arrival of news.

12:55Cameron Abadi: And apparently, it shows measurable, in fact, sort of, you know, 2x type effects on the concentration of textile jobs in those communities persisting into the 19th century. So it seems to have worked. It isn't just sort of the fabled beginning. The second phase would be the phase of the Cold War, state-backed, or the early Cold War, World War II, total war, state-backed projects of industrial and economic espionage. And they blur then just into the realms of full-on espionage. So perhaps the most successful is the German atomic scientist, Klaus Fuchs, who worked in the collective transatlantic Manhattan project and transferred key expertise.

13:32Cameron Abadi: Now, some people may know, actually, this is not a secret, but my grandfather spied for the Soviet Union on this basis and is reputed to have transferred radar information from a private contractor in the British war effort. It wasn't the well-known aerial radar, it was a gunnery radar. So that sort of thing was going on during the Cold War. Sorry, that personal connection just struck me as I was speaking. But that's literally what he did. That's quite hard to measure. We'll say a bit more about this, I think, later on, because the Soviet Union was not in economic terms terribly successful at incorporating this kind of expertise.

14:09Cameron Abadi: When we start moving into the modern period of globalization, shall we say, or the new Cold War with China, the numbers get really big really quickly. And you started off by quoting a number of 200 billion, I think. That was the loss in the capital value of NVIDIA. That's the kind of ballpark that the more hawkish American think tanks estimate the transfer of resource of IP to be from the United States to China, its antagonist. Hundreds of billions of dollars per year, they estimate. And the key element of that is not copyright or fake handbags. It's the acquisition of various types of IP. Now, those numbers form the basis for one of the more aggressive turns in American policy towards China in the 2010s.

14:57Cameron Abadi: They're highly contentious. They're clearly political. More detailed studies looking at specific firms which have reported theft rather than just kind of gross estimates suggest numbers which are not orders of magnitude but a quarter as high as that, more like$40 billion a year, plus maybe$5 billion or so in the prevention costs. which are also the sort of numbers you get out of Australian estimates. Australia is also a target, perhaps unsurprisingly, for Chinese efforts in this direction. So I don't think there's any doubt that it's going on. There's no doubt at all that it causes losses. There is, I think, a question mark over whether they're in the tens of billions or hundreds of billions.

15:37Cameron Abadi: And of course, there's also questions about maybe what goes the other way increasingly. We certainly know that American intelligence intelligence was penetrating, Huawei, for instance, was massively penetrated by American aggressive cyber warfare in the 2000s and 2010s. So, you know, unsurprisingly, it cuts both ways.

15:56Adam Tooze: Yeah, it's interesting. I mean, it does seem like one of the contentions is that there are countries, China among them, who are accused of having this economic espionage as a state policy and really pursuing it concretely. Even, you know, among other countries mentioned in that context, I recall from, you know, work at FP was France. France is also often cited as a country that has pursued economic espionage as a state policy. But the United States claims not to be doing that very vehemently. And yet the history that you were just discussing as I was reading about it as well, it seemed to suggest to me a kind of, you know, inherent irony to some extent in the history of economic espionage in the sense that the purpose of stealing these secrets seems to be to try to gain a technological advantage ultimately by incorporating them into one's own economy.

16:54Adam Tooze: But if one is successful at that, ultimately one becomes a target oneself, right? I mean, you know, we were talking about Britain and its textile technologies, but, you know, it acquired them from Italy originally before it became the target of theft of those secrets. And yeah, America, despite claiming not to pursue economic espionage, despite criminalizing in the 1990s, you know, its own history with economic espionage is pretty prolific if one digs into it, you know, in terms of pursuing British technologies in the early part of its own history and generally refusing to recognize foreign IP rights, etc.

17:30Adam Tooze: So curious what you make of whether economic espionage is sort of inherently ironic in that sense.

17:35Cameron Abadi: Yes, I mean, deep down, And after all, if one starts from anything like a kind of liberal view of economic activity, not in a normative way of this is how it should be in free markets best and even patents and infringement of that. But just if we think of economic activity as driven by like a hive mind, the competitive hive mind of interests all over the world, then the most obvious thing in the world to see is imitation, spying, you know, intelligence gathering. It's the most sincere a form of flattery after all. It's like, what you're doing is fantastic. I want to do that too. And I want a slice of the action that you're generating.

18:09Cameron Abadi: And then maybe it inspires some ideas in me as well, which then ultimately it's quite difficult to just simply reproduce somebody else's technology. This is a point we'll come back to. Successful spying always means essentially that the successful spy already has got some game going on, right? If you or I were just, you know, if you and I stumbled into Jensen Huang's bedroom in a hotel and bumped into an incredible suitcase of the latest NVIDIA chip design, I'm sure he doesn't carry them around, but just imagine it, you and I would be useless. Like it would help us in no way whatsoever. Like we have none of the requisite skills to make anything of this knowledge, right?

18:51Cameron Abadi: So successful spying always implies some capacity on the part of the spy to actually do something with it, which means you're not starting from scratch. And certainly there is this sort of back and forth, as you were saying. I believe the first, I did a thing on the newsletter about it, I think the first fully powered factory in the world was actually a British ripoff of Italian, North Italian silk textiles production, which a Brit basically more or less covertly observed, took the vision back to Britain and built a factory, which was the first fully, largely mechanized textiles factory. Yes, the US case is egregious because they were one of the world's big.

19:31Cameron Abadi: And of course, in the American case in the 18th century, and in the popular presentation, it's dressed up as patriots liberating secrets from the beastly, mercantilist British monarchy and its associated economic interests, which now in America will enable the unfolding of the freedom of America and the American Republic with much higher levels of productivity because manifest destiny. It's only in the 1890s that the American copyright actually protects not even secrets, but like, say, the works of Charles Dickens, which were systematically ripped off and produced in just, you know, what the Germans called Raubtruck, so like robbed print in the 19th century.

20:11Cameron Abadi: So it's definitely a back and forth. And in the case of China as well, you know, whilst there was still scope for a liberal narrative about China, and that, of course, now is very restricted, but not so long ago, as recently, or five or six years ago, in the pages of The Economist, you could read The Economist describing the way that perhaps the convergence story with China, at least in this limited sense, was in fact playing out. Because whereas China in the 90s, 2000s, early 2010s might have been an uninhibited acquirer, shall we say, of Western expertise and used every conceivable route to do so.

20:49Cameron Abadi: In the late 2010s, Chinese courts were actually beginning to take cases of IP theft, which included Western firms suing Chinese people, but also Chinese firms suing each other over IP. And the general prognosis was that Chinese courts, as you'd expect, would get more and more sophisticated in their arbitration. They would, in various ways, acquire a tacit, if not explicit commitment to protecting private property rights. and the story of convergence that liberals hinge their optimism with regard to China on would play out in the Chinese case. In the new era, I've heard and I've been in rooms where, and sometimes me being me, I like sort of, you know, cynically and somewhat flippantly introduce this idea, but I don't ever get any dissent, which is, isn't the idea now with regard, say, to Chinese EV factories, electric vehicle factories, that we get them to build them in Europe.

21:42Cameron Abadi: And then the phrase that people use is, this time we'll do to them what they did to us. In other words, we will acquire technology from them quite rapidly, whether it's batteries or chips. But it's almost always about electric vehicles, because it's quite clear that without that, even the most sophisticated European producers at this point are in trouble. I mean, they actually need to organize a reverse transfer. And the Chinese anticipating this have actually also moved very aggressively towards criminalizing the transfer, which of course puts any Chinese, given the sanctions the Chinese state can impose on people, that puts Chinese engineers in an incredibly dangerous position if they're in joint ventures.

22:22Cameron Abadi: So Beijing can see Europe coming in manipulating this gradient. But given history, you wouldn't bet against a future in which in 10 years from now, there are firms in Central Europe and Europe which are basing their business model on technology that flowed the other way.

22:42Adam Tooze: Yeah, this raises for me the question of methods when it comes to economic espionage. I mean, you know, in one sense, if you had the acquisition of these secrets or maybe knowledge more generally is whether espionage is just one variety of acquiring such knowledge. You know, when we talk about espionage, some methods that come to mind are, you know, human intelligence of the kinds that involve using spies to acquire secrets. And that seems to be what is at stake in the NVIDIA story right now. There is cyber intelligence to acquire secrets using digital methods, signals intelligence. But all of that seems to me contrasted with some kind of like commercial and legal methods of pursuing espionage, so to speak?

23:33Adam Tooze: I mean, like hiring of engineers from companies, say, or acquiring companies that are at the leading edge of certain technologies. I mean, when it comes to these methods, which are sort of most effective or how do they sort of contrast with one another?

23:49Cameron Abadi: Well, the ones that we know systematically work are clearly acquisition, joint venture. That's been a channel for the transfer of technology for as long as there've been those kind of projects and many of them actually have it as part of their terms it's part of the deal that you know x amount of expertise will be transferred so and so during the the post-colonial period there was an effort to so-called quote-unquote nativize the workforce which is train up people who are not expat western europeans or whatever so that they can they can take over the running of the firm in due course saudi aramco would probably be the most dramatic instance of this worldwide, but then Chinese industrialization is heavily driven by this kind of process.

24:31Cameron Abadi: So I think that's the one about which we can say without any kind of ambiguity that it clearly is a powerhouse of economic development. And the liberal in me says, bring it on and let's please, please, please not criminalize this. This is absurd, right? This is actually just development. Down the other end of the spectrum, you know, there is everything from signals, intelligence, large-scale monitoring and cracking of commercial secret information. And then people and pressure on people, communication between people, everything from people being paid to transfer blueprints of things to casual pillow talk.

25:13Cameron Abadi: Presumably, the risk with Jensen Huang is not that he shows up somewhere in a hotel room with blueprints, But I don't know that his personal communication is so completely compromised that hackers can get in. So presumably in the modern age, like the way this works is that somebody gets entangled in a relationship like this. it turns out that it's a honey trap of some kind and then either blackmail is applied or or it's just a doorway into their electronic world and then you can start transferring the terabytes of data which it presumably you know you can't i was thinking to myself i don't think adam you can actually put this in a suitcase anymore hey you know it's not that kind of intelligence we're not in the 1980s anymore so i would have thought in the modern age all of these things go very much together.

26:04Cameron Abadi: But to go back to the sort of normal end of this, it's all about things like non-compete clauses. How long is it that you have to wait before you move from one tech business to another? This applies to the financial sector as well. People have to do, they go on gardening leave for a period of time, so-called, before they're able to switch, say, from one trading team at a big investment bank to a hedge fund or whatever. And there's actually quite a lot of interesting studies about the very different cultures that are created, depending on what type of non-complete, non-disclosure rules actually stand up in different jurisdictions.

26:44Cameron Abadi: So there's differences, for instance, between East Coast, Massachusetts law on this and Californian law. And there is apparently a body of thought that says that one of the reasons for California's and Silicon Valley's enduring competitive advantage is that Californian law is actually restrictive with regard to non-compete clauses. So it's relatively easy for human intelligence to move, as opposed to Massachusetts, apparently, where it's more vertical. In other words, it's harder for bodies and minds to move. And the crucial thing about that, to just come back to absolute basics, is what is it that you acquire?

27:18Cameron Abadi: If you get a blueprint, you've got a blueprint. If you've got a flow of communication, you know, I don't know, some huge cyber operation, then you have to crack that and make sense of it and run algos on it to figure out what it is. But if you've got an actual person, they embody all sorts of tacit knowledge. So it's really a question then of putting them in your lab and seeing what they do, because they may not even actually be able to spell out to you all of the different requirements that are necessary to make a particular experiment work or replicate a particular result in a blueprint form.

27:55Cameron Abadi: And so I think that's the kind of spectrum of various options here.

28:01Adam Tooze: Yeah, I mean, it's interesting to reflect on this conversation as we're having it and how easily we're treating the kind of like legal methods and the, you know, ostensibly illegal methods that's existing on a spectrum is surprising on some level, even if functionally, yeah, they are obviously the same. Am I not being judgmental enough? Sorry. No, I'm sorry. It's interesting from, you know, you're citing your own, you know, liberal bona fides. And it's interesting, you know, it always seems to me, you know, another liberal perspective would be to emphasize the importance and the virtues of consistent rules.

28:36Adam Tooze: But of course, as we know, there's also a lot of hypocrisy always whenever rules are in play and who sets them and breaks them, etc.

28:43Cameron Abadi: And particularly with regard to knowledge, right? Because knowledge is a non-exclusionary type of good, right? me having it doesn't mean you don't have it. And there's a fundamental tension in liberal thinking about patents that on the face of it, they're an abomination. They're basically a licensed monopoly. And the strongest argument for them is that otherwise, you know, people don't have the incentive to do the hard work of innovating. And they don't have the incentive to make the investments which are necessary in modern types of scientific endeavor and engineering endeavor. But they're deeply compromised.

29:17Cameron Abadi: And that's also the argument for, and I'm not arguing the case now for my grandfather, but that's also the case after all the argument for the relatively lenient treatment of spies, certainly during World War II itself, because the information was being passed to allies. You know, why is it that it makes sense to monopolize the scientific knowledge of Western allies when the vast majority of the fighting is being done by the Soviets? And that's obviously a rationalization. And it was clearly, it was, I think, one of the key reasons why people who were exposed during the war itself were treated under a very different regime than folks during the Cold War.

29:51Cameron Abadi: So these logics are complex, I think. And ironically, one of the weirdest things is that if you don't want to disclose anything, you don't ever file a patent. So it never actually gets fully classified as intellectual property. So then it falls under a different type of law, which is much harder to actually classify, which is business secrets. which are by definition not formalized, right? So the more you formalize it, the more you actually expose it. And so then ultimately the least disclosed form of knowledge is in the body of people. And non-compete clauses are pretty serious infringements of people's civil liberties and can be pushed in a very aggressive way in modern American law, corporate law, right?

30:29Cameron Abadi: Non-disclosure, non-compete clauses can materially inhibit people's freedom of movement between jobs.

30:35Adam Tooze: We're going to take a quick break right here, but we'll be back in a second to continue talking about economic espionage.

31:07Adam Tooze: all 13 ,775 feet of it. Now in addition to making sure we had the right ropes, helmets, and footwear we also had to ensure we had the right outerwear. So that's when I turned to Steeo. Locals in Jackson know Steeo's the brand climbers, skiers, and trekkers wear because Steeo is local too. Their gear is designed and tested right there in Jackson Hole right at the base of the Tetons. The hometown down jacket is 800 filled down, but feels more like your favorite hoodie than a puffy. And the environment jacket, pants, and bib are your all-mountain workhorse. Now, I know when I hit the summit, it felt amazing to be there bundled up warm and feeling gratitude toward my friends and guides.

31:49Adam Tooze: And now, for being loyal listeners, you too can feel the difference this winter and gear up the way locals do. Go to stio.com slash ones for 20 % off your order. That's S-T-I-O dot com slash ones. Stio dot com slash ones. Stio. Designed and tested in Jackson Hole.

32:10Cameron Abadi: Hi, I'm Laura Ross-Broutellum.

32:12Adam Tooze: I'm a producer here at Foreign Policy. And I'm here to tell you that this episode is sponsored by BetterHelp.

32:21Cameron Abadi: If you could permanently delete just one thing from your brain, what would it be? I know for me personally, my house and its mess just occupies way too much space in my mind.

32:35Adam Tooze: Just my inability to keep things organized, it takes up a lot of my time. And I feel like I'd be much happier if I had the time and the strategies to deal with it. And honestly, that's something that if it's bothering you too, or you have something else that just takes up a lot of space and you can't get off your back, Therapy could be really helpful. BetterHelp therapists are licensed, experienced professionals, and BetterHelp does the initial matching work for you, so you can focus on your therapy goals, whatever they might be. BetterHelp is now in network with more than 125 health plans, with many co-pays as low as$15 for eligible members.

33:20Adam Tooze: And if you don't have eligible insurance or you live outside the United States, you can still get 20 % off your first month. Sign up and get 20 % off at betterhelp.com slash ones twos. That's betterhelp.com slash ones twos. To return to this conversation of methods, you know, you mentioned honey trap or honey pot espionage as a method. And I guess more generally, yeah, that involves creating a sexual relationship in order to exploit that relationship in various ways. whether to gain access or create blackmail, etc. And, you know, without wanting to indulge in stereotypes, but it does seem like it's fair to raise the question of whether the relevant economic actors in our, you know, tech economy now are particularly vulnerable to that method.

34:16Adam Tooze: I mean, we're talking about a kind of tech world where, you know, people have risen to the top by being kinds of really obsessive, focused, nerdy kinds of people, maybe? Is it fair to wonder whether honeypot operations of the kind that have now been exposed perhaps at NVIDIA are a particular problem now?

34:36Cameron Abadi: Yeah, I was thinking about this as well. I mean, it's almost impossible. I'm not entirely going to resist the temptation to speculate about, you know, the ego makeup of Silicon Valley engineers. It wouldn't, after all, be surprising if there was a desire for compensation and for overcoming a whole variety of inhibitions, which no doubt are channeled at various stages in many people's careers into extraordinarily efforts in very, very dry, very recondite zones of technical expertise. So, you know, the nerd gone wild, the head of a company, the loneliness of people in that kind of position, all of this, I think, no doubt feeds in.

35:15Cameron Abadi: But then I was thinking actually sociologically, and I came to this strange realization because I spend quite a lot of time in hotels because of this weird existence that I have. And I was thinking to myself, if a senior soldier goes to a hotel, on the face of it, you're kind of wondering what the senior soldier is doing in a hotel, right? If a senior civil servant, pretty much whichever government they work for, goes to a really glamorous hotel, again, you'd be asking yourself, how can they afford that? That's not really where they belong, right? So suddenly you realize that the whole zone of high-end, expensive, relatively exclusive hospitality that presumably provides the most attractive zone for this kind of entrapment, the actors for whom, where it arouses zero legitimacy issues, zero need for explanation at all, in fact, it's completely normal and taken for granted, are high-level business figures.

36:14Cameron Abadi: You know, the entire conference circuit where people stay, the blending in those spaces of leisure and business activity, including pools and gyms and eating and drinking as a matter, of course, in fact, a kind of mandate for many businesses. I go to a lot of these investor conferences and it's quite clear that networking is an absolutely required option. There's no excuse for just withdrawing to your room and getting on with your life, which I'd want to do most of the time. And so presumably this is precisely the world in which this works. I mean, business people systematically expect to be wined and dined.

36:52Cameron Abadi: They expect to be courted. They expect in various, let me put this, they constantly struggle to draw the line between reasonable inducements and bribery. That is an essential part of their business. And many of them, of course, have very strict codes of ethics, but it's a matter of record that those shift all the time. and they're much more like in that sense the Chinese Communist Party than they are like conventional civil service jobs in the West where it's just quite clear that accepting any kind of gift in any form is just ultra-vires. No one in the civil service is regularly invited to glamorous parties where there's a mixture of more or less formally clad, in some cases scantily clad, different types of interaction.

37:34Cameron Abadi: And so it's a kind of recipe for stress on this fault line, I would have thought. And it's purely sociological. It's not a matter of psychology at all, though the psychology that then is acted out in those spaces is, of course, quite interesting in the games that are played around, you know, dress code and stuff like that at really quite a high level. So that that I think might be one way of rationalizing this, that if you were going to target those kind of actors, this would be an obvious way to go. Though it's worth saying that I think some of the most successful honey trap projects were run by the Stasi in West Germany.

38:10Cameron Abadi: And they were actually quite often targeted at civil servants. And it was quite often women. So some of the most significant cases, the Stasi would call them Romeos. And their job was to target West German women researchers, female researchers who were interested in East Germany. and the allure was in part erotic and part just intellectual. You know, people wanted to talk. And then they were highly successful, in fact, at placing people who were suborned in this way into the Bay and Day and into the German intelligence and senior ranks of the German civil service. So it's quite, you know, it takes a different type of structure.

38:49Cameron Abadi: The more common thing that we seem to be dealing with here, I think, does fit into this sociology of business sociability that just creates many more risks and hazards of this type than is commonplace, especially for civil servants in general, I think. And military in particular, because it really would attract attention. I think it would be an obvious red flag if you saw a senior military figure enjoying hospitality in high-end restaurants or hotels anywhere in Asia.

39:19Adam Tooze: Yeah, I think the sociological element of this is necessary if we're thinking about the kinds of inducements that are being offered. And yeah, the assumption in describing something as a honey trap is the suggestion that the sex is the main inducement and that's really what is being offered. But obviously these things are always more complicated. And yeah, in the East German case, you know, where the genders are switched, I think it's usually described as it wasn't the sexuality so much as like, yeah the kind of sort of like companionship or loneliness being addressed by by a romantic figure i mean who knows i i'm not uh i'm steeped in the in the details of that history but uh yeah one

40:01Cameron Abadi: only has to read like one only has to read you know serious by-fetching like john le caray or people like that to realize that clearly intimacy is the biggest no this is a matter of seduction fundamentally so what is it and in each particular case it'll depend on what it is that people need to be, you know, what is the trigger and presumably skillful spy masters or whatever, just are very good at finding the lever through which to pressure people and through which to seduce them. And it could presumably be sex, but it much more likely, I would have thought in general, is something more about recognition and attention.

40:36Cameron Abadi: And then ultimately, some kind of, that would seem to be a much more plausible story. Because these spaces that we're talking about sociologically are also the spaces of high-end escorts. So if what you want is sex, there are many, many ways of obtaining this, especially for somebody with enormous wealth that do not require, I mean, all of them involve certain amounts of risk, but certainly don't require, you know, this kind of complexity. I mean, somebody like Jensen Huang, if they're interested in sex, presumably can get it in a much less complicated way than this.

41:11Adam Tooze: So I do want to return to a distinction you were mentioning, drawing earlier in this discussion, and that's between pursuing the acquisition of technologies through espionage of this kind that we're talking about versus productively absorbing that technology upon having acquired it. And how relevant is that distinction? How does it work and how does it apply right now, say, to China? I mean, would or would not China be effective right now at implementing the secrets it's being said to be trying to steal?

41:46Cameron Abadi: I think that's the whole game, actually. The real reason to be worried is that China has a demonstrated and extraordinary capacity to incorporate. In other words, it's a near peer. and that's precisely what makes it so alarming to Western security officials is that, you know, there's absolutely every reason to think they can, yeah. And furthermore, leapfrog even, like give them half a chance and there's every single reason to think that they can generalize it, make it cheaper, speed it up. There are differences, obviously, because what NVIDIA does or what TSMC does or what, you know, the lithograph makers do is literally special source.

42:25Cameron Abadi: Like no one has ever done that anywhere ever in the entire world in the history of humanity ever. So whether you can replicate that is an open question. But that's presumably what the Chinese aren't going to do anyway. They're going to make their own version that's more suitable to their circumstances. And it's critical to know. And we don't know what's actually happened in this case, but there were certain critical problems that it's interesting to know how other people solve them because it helps you to think about how you should solve them in your own terms. And that is why I think the Chinese case is so significant, because they have a demonstrated capacity to do that.

42:58Cameron Abadi: And it really varies, like in regard to transport engineering, cars, railways, solar, batteries, hugely successful, leapfrogging, game is over. With regards to chips, slightly different story. And I think that speaks to the fact that the technological demands are just so mind-blowingly high. And I think something similar is also true for jet engines for the very latest generation of aircraft. The Chinese have been acquiring that technology from the Russians who are basing their technological development on the Soviet era. But there are some structural issues with their ability. the last time I looked anyway that was regarded as a key issue and acquiring and sourcing information from the weak spots in the former Warsaw Pact which were Russia and Ukraine was seen as a way in which the Chinese might able to leapfrog and now of course perhaps with a new era of drones none of it matters very much anyway and they're very apparently quite advanced in rocketry so maybe that's an old story but that does contrast very significantly with the Soviet Union, which at different periods in its history demonstrated greater and lesser degree capacity to incorporate.

44:08Cameron Abadi: So around the atomic bomb program, quite a high capacity to incorporate. Then, of course, in rocketry, they become world leading. And they do that like the Americans by capturing German rocket scientists and then incorporating their knowledge into what was already a high functioning military industrial system in the Soviet Union that was wasteful of people's lives, but was generative of a whole series of incredibly innovative, very, very powerful and effective weapons. So it's not as though the Soviets need this to suddenly be a world-beating power, for heaven's sake. They beat the Wehrmacht in a knockdown, drag out fight.

44:43Cameron Abadi: But yes, when they then capture German rocket scientists from the V1 and the V2 program, they know what to do with them. But more generally, it's a remarkable instance of just how little can get transferred, because after all, the Soviet Union was the beneficiary after 1945, of a huge wave. This is the untold story of German reparations after World War II. Absolutely gigantic. And they take the form of dismantling. So vast amounts of German equipment is shipped to the Soviet Union, where I don't think it's merely apocryphal to say an awful lot of it just rusted and was not fully incorporated into the Soviet system.

45:17Cameron Abadi: Now, in earlier stages, they had done much better. So in the 30s, the Stalinist industrialization program is heavily dependent on American expertise and on German expertise funded in part by export guarantees from the Weimar Republic that helped to fund this flow. So it wasn't consistently rubbish incorporating, but it certainly wasn't the powerhouse of creative imitation, adaptation, modification, improvement that China is to an extent that nowhere else in the world can match, right? That's the extraordinary thing. They can iterate, they can modify, they can incorporate at pace that no one else in the world.

45:53Cameron Abadi: In almost all areas of production, no one else can match. And so then if you feed that with privileged insider information, then you'd expect really large-scale effects. But of course, that implies a huge capacity on the Chinese side to start with. In other words, this is not a story in which it's only by way of the acquisition of the foreign expertise that things happen.

46:15Adam Tooze: I do want to briefly emphasize that image of Yeah, the Soviet Union literally dismantling factories in Germany, moving them all the way to Russia and then just having them sit there. It's quite a wild image.

46:27Cameron Abadi: Because the Chinese have dismantled factories in Germany as well. But we know for damn sure those are operating and those were used as the template for the build out of some of their heavy industrial capacity.

46:37Adam Tooze: But to take a closer look at this case, I mean, how significant is this alleged espionage with NVIDIA or more broadly with these, you know, USAI and tech companies? I mean, is the, you know, allegation that China is acquiring chips that it can then use to, yeah, leapfrog the US or even like reverse engineer these chips and things that it's maybe acquiring from these companies? I mean, how significant is what's going on and what would that be exactly?

47:12Cameron Abadi: I mean, we just literally don't know. All we know, as you were saying at the beginning, what we know is that NVIDIA is the most valuable company in the world ever. it's the hub of the AI boom for loads of reasons we've talked about on the show before and keep thinking back to the fact that you had the sense to say Adam we should do a segment about NVIDIA years ago now and that was the first time I'd ever thought about it and since then it's become literally the most the most valuable company in the world in history and what seven percent or eight percent of the S &P 500 on a good day and the anchor of the entire internally self-financed AI system.

47:47Cameron Abadi: So it's hugely important business. But of course, it's full of brilliant engineers like Jensen Huang, maybe like the driving inspiration of this, but he's not the man that actually does everything. So who knows how important it is that he's having this affair. We know that there've been allegations that Alice Huang is a MSS asset, but they flat out denied it. The firm that she's involved with, Megaspeed, which is one of these crossover companies, right? It comes out of gaming and has moved into data centers. It's a little bit like the crypto miners that have moved into data centers. And NVIDIA itself comes out of that gaming zone.

48:25Cameron Abadi: It's not even on the entity list, right? So, I mean, there's been a push apparently within Washington for it to be put on that list. But like many part of the universities of the Columbia University on the entity list, it's not even on that list. And it doesn't seem to have been a very substantial business. So who knows what this actually adds up to I mean, I have to say that it's hard not to see this, to go back to the sex theme, as part of the kind of liberal American nightmare, right, of inappropriate sex, outside marriage, in a corporate setting, conflicts of interest, and then also Chinese and AI.

49:03Cameron Abadi: And then, of course, all of the alarm bells are going off everywhere and nothing actually needs to be proven. I mean, what's so remarkable is this whole thing explodes with it's all smoke and so far zero fire. But the sense that this is the nightmare is so palpable that the very least sign that it's arrived is immediately, quite obviously, whatever it is. And the rumor mill in D.C. gets going. And presumably, The New York Times has some substance to the story. There must be some people making some claims which are halfway credible. But how much one can trust any of those? To be honest, who are we to say at this point?

49:40Cameron Abadi: Is it plausible that the Chinese should want to do this? if I was the Chinese spymaster, would I be trying to do this? Of course. So if Jensen Huang is out in the open socializing the way he does, you try and suborn the guy, obviously. But who knows? I mean, what's remarkable about him is how freely he's traveled around the world, how accessible he appears to be, how much he likes socializing. What a rock star he is, right? The guy's a rock star and he acts the rock star and he looks great. I mean, it's not like this is some sort of hardship mission seducing him. Like, I don't know. The whole thing has about it, like the chronicle of a death foretold or the chronicle of a disaster foretold.

50:21Cameron Abadi: It's like, what, he doesn't have a Chinese spy girlfriend yet? There's something slightly wrong here if that isn't the case. And so now it's arrived and we all go, oh my God. And yet we don't actually have any, as far as I know, the FT summary of it basically just says we know NVIDIA is important and he's accepted that there was an affair. or he's acknowledged that there's an affair. But what else? And she doesn't seem to have made any secret of it. That's the incredible puzzle. If she was actually a spy, why does she have it all over her social media? Because I've seen pictures of them together for months, way before this blew up.

50:56Cameron Abadi: I'm sure millions of other people have also seen these pictures of him, consistently with the same person over and over again. And she doesn't make any secret of it. So maybe this is double think that if she did keep it clandestine, that would look more suspicious. But the whole thing about it's so obvious, if you like, that it's difficult to know how you negotiate the clandestine and the surprising in a world in which this just seems as predictable as well as anything could possibly be.

51:24Adam Tooze: Yeah, I see what you mean about it being overdetermined. Somehow it begins with with his leather jacket.

51:29Cameron Abadi: He looks fantastic. And he like jumps on stage in these crazy things. And And there's all these cheering fans, including many, many extremely eligible women who are going, oh, my God, oh, my God, you're so amazing. And so it looks more like a Beatles concert or something like, you know, Jensen mania.

51:45Adam Tooze: I mean, I say in jest, but one can wonder whether it's actually a U.S. intelligence operation, maybe.

51:51Cameron Abadi: Well, there have been people saying that, like, you know, this is all associated with the midterms and the AI bubble that was bound to bust. And apparently they've known about it for months and months. right? Apparently Trump had reports on his desk in May. And Lutnik maybe at Commerce is blocking putting this company on the list. That's what the FT reported or Lutnik's in charge of Commerce and no one understands why Commerce hasn't taken the action. So it all just seems both totally obvious out in the open and yet actually kind of vaguely obscure.

52:23Adam Tooze: Well, I guess this relates to the last question I wanted to ask, which is how economic espionage potentially relates to counter espionage. I mean, does espionage of this kind expose vulnerabilities ultimately to forms of response like sabotage? And I'm, you know, thinking of certain parts of Cold War history where there were countermeasures by the West against the Soviet Union. And yeah, how does that history unfold exactly.

52:55Cameron Abadi: Yeah, this has been a fun episode to research because apparently there was an amazing, I imagine you're talking about this episode, the farewell, Operation Farewell, where there was a defector, Vladimir Vetrov, a KGB officer who had the good taste to defect to the French. So he's having a good time in Paris. He takes a dossier of material to the Mitterrand government in 81. And he's from the T office within the KGB. So this is the technological spy network. And they were running assets all over the West, thousands and thousands of people, apparently. They were known as ex-officers. And so he brings in this network and says, look, this is what we're doing.

53:33Cameron Abadi: And rather than panicking, the Western side apparently decides to just play this back on the Soviets. So because there's so much information flowing, presumably it's quite difficult for them to decide what's plausible and what not. But they start feeding the Soviet broken things, basically, or things which were themselves then weaknesses, and variety of bad chips and software that doesn't really work properly or industrial machinity, defective plans. And there's even a rumor that there was a really nasty explosion on the Siberian gas pipeline years later that may have had to do with the fact that this operation planted defective buggy software in the Soviet than the Russian energy system.

54:18Cameron Abadi: That's highly disputed and debated amongst people who are into intelligence history as to whether or not that actually happened. But it's undisputed, I think, that seeing this interest from the Soviet side, the West were able to exploit that against the Soviet side by planting misinformation. so yeah there's all that kind of options for these sorts of cat and mouse games and this just brings me back to this basic question of getting what does getting jensen hong into bed actually do for you in terms of winning the chip race i i'm not entirely i'm not entirely certain you know when people ask me whether i think i'm under surveillance from mss or whatever i always say well they're really wasting their time in that case i'm part of the counterintelligence operation like what what is it that what is it that you think you could possibly gain that would be of any use from my brain like how would that how would that help your situation so i just think there's a tendency to kind of dramatically rationalize this as um you know targeted highly strategic massively successful influential and he may as we've said like there are clearly incidents in history, it really did matter.

55:31Cameron Abadi: But on the face of it, it's, yeah, we would need a very complex understanding of what's actually been transacted here before jumping to conclusions, I think.

55:41Adam Tooze: Well, I'd like to think that spy services would like to know how we're preparing for this podcast. You know, I don't know.

55:47Cameron Abadi: Well, this is the thing, it's so narcissistic. Like there were these headlines, these headlines in the British press that I was mocking on Twitter, which was like, who's next? It's like, you know, Who gets to be seduced by MSS next? It's literally sort of because I'm so precious and so important. And so surely they're going to want to send some glamorous spy after me. It's at some level, it's such an x-ray through the mind of the West.

56:12Adam Tooze: Well, if any spies out there want to know, just write us an email. We'll include you on our emails. But otherwise... Come to the live show. Exactly. Come to the live show.

56:24Cameron Abadi: It's a classic rendezvous for this kind of thing. Yeah.

56:27Adam Tooze: exactly uh but for now we should say goodbye and of course we'll be back next week with new data ones and twos is written and edited by adam and me and produced by claudia tady the show is made possible through the support of foreign policy readers and ones and twos listeners as always we love getting listener feedback you can leave voice messages on the ones and twos homepage on foreignpolicy.com or email us. That's podcasts at foreignpolicy.com. Or you can tweet us. That's at onesandtwospod. I'm Cameron Abadi. Thank you very much for listening. And we'll be back in your feed next week.

57:23Adam Tooze: Are you ever shopping online and suddenly a discount that helps you save pops up? It makes your day, sometimes even your week. Want that feeling month after month? Well, new customers who save by switching their home and auto insurance to Progressive save hundreds. Cue your happy dance. Get your quote today at Progressive.com. Progressive Casualty Insurance Company affiliates and other insurers. Not available in all states.