China Is Undercutting America’s AI Giants

0:00Support for the show comes from VCX, the public ticker for private tech. The U.S. stock market started history's greatest wave of wealth creation, from factory workers in Detroit to farmers in Omaha. Anyone can own a piece of the great American companies. But today, our most innovative companies are staying private longer, which means everyday Americans are missing out. Until now. Introducing VCX, a public ticker for private tech, now available wherever you buy stocks. Visit GetVCX.com for more info. That's getvcx.com. Carefully consider the investment material before investing, including objectives, risks, charges, and expenses.

0:36This and other information can be found in the fund's prospectus at getvcx.com. This is a paid sponsorship.

0:46This episode is brought to you by Accenture. When your advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales. using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business. The result? Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more at Accenture.com slash Spotify. This episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome, that's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50-page restoration block, or finally break down that long article you've had open for weeks.

1:31Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses set up required, compatibility and availability varies 18+.

1:47Money market's mad. If money is evil, then that building is hell. The show goes on! The folks are never watched. Welcome to Prof G Markets. I'm Ed Elson. It is July 28th. Let's check in on yesterday's market vitals. The major indices were mixed as chip stocks sold off. Brent crude fell below$90 per barrel as the US and Iran paused their attacks. The yield on 10-year treasuries declined. Chinese chipmaker CXMT soared more than 400 % in its public debut, becoming China's new most valuable company. More on that later. And finally, Apple rose more than 1 % to overtake NVIDIA as the world's most valuable company.

2:32Okay, what else is happening? Chinese open models are gaining ground and Washington and Silicon Valley are starting to notice. In recent weeks, Chinese startups ZAI and Moonshot AI have released models that are competitive with those from US Frontier Labs, the top five most popular models on OpenRouter, a marketplace that tracks usage across AI models, are indeed Chinese. And when a rogue, unreleased OpenAI model hacked HuggingFace during a security test last week, HuggingFace used a Chinese open-weight model to defend itself. Now, you might have thought that American AI executives would use this opportunity to encourage a crackdown on Chinese open-weight models.

3:13Instead, the opposite happened. In his first ever post on X, Jensen Huang shared a letter titled, Open Weights and American AI Leadership. And within a couple of days, 50 companies signed the letter, including Microsoft, Meta, Palantir, and IBM. Together, they argue that US AI leadership depends on building not just frontier systems, but a strong open ecosystem around it. And they warn Washington against, quote, premature restrictions that stifle competition and drive innovation overseas. Which raises an interesting question. And that is, why would US AI leaders seek to protect the very thing that supposedly might destroy them?

3:52Joining us to discuss this, we're speaking with Scott Singer, Technology and International Affairs Fellow at the Carnegie Endowment for International Peace. Scott, thank you for joining us. I'd love to just start with the context here. This open versus closed wait debate that has taken over the world of AI has a lot to do with U.S. versus China, which is strange and interesting, but I would appreciate if you could just clarify why that is and how we got to this moment. I think it's really a lot of happenstance more than it was any initial intentional strategy or ideology. If we look at the history of AI model development, a lot of the initial general purpose frontier AI models that were coming out of places like OpenAI and then Anthropic were proprietary.

4:43The weights were not accessible to any sort of outside users. And so for China, there was a question of where and how they might be competitive on the global marketplace for advanced AI models. And so having models that, for the weights, that are openly available and can be used by companies and startups and can be fine-tuned or adjusted so that they can fit whatever purpose they have, that was sort of the area where, you know, China couldn't necessarily compete on having the most advanced capabilities, but they could have leaner, cheaper models that would be useful across the startup ecosystem and for companies.

5:19One of the things that has caused a lot of controversy is this idea of distillation. And this is what Anthropic and OpenAI have accused many Chinese AI companies of doing, distilling their models, which has a relationship between being an open weight model as well. what is distillation and why does it matter in this conversation distillation is basically taking a much stronger model or a bigger model and using it to build a sort of smaller cheaper model and so that is you know the sort of crux of what's happening here and you can in an open model you can take in these cases Chinese companies but there's been allegations that American companies doing this as well um you know taking what is american innovation american products and using that to power these chinese products and distillation is a super common practice in ai and on its own is not sort of something to you know be too focused on or see as especially malicious but the question is really tied to intellectual property uh trade secrets as well is fraud.

6:26So the way in scale that some of the Chinese companies have been accused of sort of leveraging American IP in this case is through, for example, building up a bunch of fraudulent accounts, getting the outputs from the American AI models, and then using that output to train their own models. And so there's this sort of thorny legal battle that gets at the heart of a sort of broader geopolitical question, which is basically, you know, what are the rules of the game in terms of this very technical process. And it is more representative of this sort of open versus closed battle, even if distillation itself is sort of a widely accepted common practice across AI.

7:08I mean, just from the perspective of a Chinese company, Anthropit releases its model and then you submit like thousands of prompts to fine tune that model and then eventually replicate that same model. And that is what many of these companies have accused companies like Kimi and the Kimi K3 model, which went viral and made a lot of headlines last week. That's what a lot of these Chinese companies are being accused of. Now, I mean, in the rules of, you know, Claude and Anthropics privacy policy, you're not allowed to do that. I guess the question for Chinese companies is, do they care about that?

7:44The answer is probably no. So then it seemed like we were going to get some regulation because the administration seemed to have a view on this, and they were accusing these Chinese companies of distillation and saying that that was a problem, which is why I was quite surprised to see this open letter from Jensen Huang, co-signed by a lot of other American companies saying, no, don't restrict this. Let this happen. Distilling is okay. Open source is okay. Open weight is okay. Why do you think they are down with this? I think that there's maybe two things to point out. The first is that the American tech companies are not a monolith and nvidia itself benefits tremendously from you know exporting many chips to china it can power chinese air development and much of the american tech stack remains even in the era of sort of great firewall tech decoupling so much of u.s tech is still integrated deeply with Chinese tech.

8:43And so I think in that sense, it is beneficial to companies like NVIDIA to make sure that open way to access remains open. Startups in Silicon Valley, the A16Zs of the world, are running on a combination of American and Chinese models. And so American startups actually want to use a lot of these Chinese models because they're so malleable and adjustable. And so I would say those are sort of the main factors that are really driving this sort of bifurcation between, on the one hand, you want, you know, the anthropics of the world that really have not much to gain and are getting their profits trimmed off by the fact they have to face competition from the smaller Chinese companies.

9:30But then you have so much of the U.S. that still benefits. Yeah, it seems as though this might be kind of an attack in so many ways on OpenAI and Anthropic, because, I mean, those are the two companies that really lose if the Chinese cheaper open-weight models continue to gain market share because it puts pricing pressure on them. Might that have a role to play in why someone like, you know, Mark Zuckerberg at Meta would be very excited about rolling back any restrictions on Chinese companies? Because essentially it means that you kind of slow the roll of Sam Altman and Dario Amadei. Is that kind of, are those the teams that are emerging in AI right now?

10:14I think it's undoubtedly the case that for a company like Anthropic, which is exclusively building these proprietary closed source models, that having really strong competitive open source models is not really where you want to be in terms of, you know, you already have razor thin expectations in terms of the data center build out. You're investing a ton in compute. And if you have your profit potentially undercut, you definitely don't want that. I think that there is just sort of a question for the entire ecosystem of how far is too far when it comes to restrictions. And I think that in general, there's going to need to be a complementary approach that includes a combination of closed and open models.

11:00We see from, you know, instances like the OpenAI hugging face incident last week that there are going to be some serious risks, including the possibility that, you know, there are serious cyber incidents that lead to loss of developer control of varying degrees. And we also have concerns around misuse. And a combination of open and closed response are probably going to be necessary to mediate the full range of concerns that we would have coming from those models. I think the other thing this brings up is just the rise of China in the AI race. And I know this is something that you spend a lot of time focusing on.

11:39I mean, where are we in that race? Where are we in terms of Chinese investment in AI, Chinese development AI, and then also Chinese adoption of AI just among the general populace? The numbers in China in terms of investment are not what they are or close to it. in the U.S., but we do see a very powerful fast-following strategy where China's not really trying to be at the frontier of capabilities or the most prominent of where, or the most elite of where U.S. model providers are, but they want to stay a few months behind it. If they can be maybe three, six, or nine months behind, then perhaps that's not really going to make or break their competitiveness, especially if they continue to have access to U.S.

12:29models through distillation. And so what does this mean for China? It means that they have to operate and get the best they can within an environment where they don't have that many financial resources compared to the U.S. companies, where they have far less access to hardware to train on. You know, they're making the best of a strategy that has been sort of forced on them. And so I would bet on the U.S. position overall. You know, the U.S. is the model provider of choice across, you know, OpenAI, Google, Meta, Anthropic, XAI. Like, if you look at the global diffusion of these models, there's just not that many countries and companies that are really uptaking outside of the startups that are really excited about the Chinese ecosystem.

13:17America is still really dominating here. But I think the question is just how far can China compete outside of its own borders? And one really interesting indicator of this is, like, for example, at the launch earlier this month of the World AI Cooperation Organization in China. And if you look at the list of countries that signed on to the list for whatever this international cooperation organization is going to be, it's not exactly a really big, powerful group. You see absent on this list, you know, really none of the countries in the Middle East that have been pretty excited about Chinese investment.

13:52And so it is to say that amidst this moment where Kimi K3 and Moonshot are getting deserved attention for the legitimate, really strong capabilities that their models have, there is just like an overarching gap between how strong the U.S. is, I would argue, and how strong the Chinese ecosystem is. This seems to be one of the biggest concerns in the AI world right now is, is China going to beat us? Do you think that investors are too concerned about China, more concerned than they should be? I think it depends on what exactly you're concerned about. I think if you're concerned about, you know, potential losing profits because open source is a few months behind, I think that that's a legitimate concern.

14:35I would still bet if you are betting on whether or not the U.S. is going to have extremely capable general purpose capabilities first and that that is what is going to power economic and strategic advantage in the long run, then I would wholeheartedly bet on the U.S. ecosystem. But I think it's really at the profit margin level. It's, you know, what models are startups building on? And I think that there is a question in terms of the battle for the rest of the world. If you're sort of in the world post-worldly cooperation organization where most of the world just thinks that U.S. models are better and more capable, then that seems like a fine world for the U.S.

15:12But it's an open question if, you know, China begins to more seriously aggregate up its compute to make a stronger play at general purpose capabilities in a more aggressive way than it is currently. So I think let's pay attention to market consolidation on the Chinese side. Let's pay attention to just how far behind they are on capabilities and how successful they are on diffusion. This has been a massive priority for the Chinese. Last year, they launched the AI Plus strategy, which is basically focused on embedding AI into practical, useful applications. And so far, it's really not clear exactly how successful this strategy is going to be.

15:54It seems like it's really sector-specific around how easy it is to embed AI into your ecosystem. In the U.S., of course, it does this in some ways just through market pressure. We see our own economy fundamentally transformed by AI capabilities. And so even if it's less government-driven, I think that both societies are being dramatically transformed by the rapid adoption of AI. All right. Scott Singer, Technology and International Affairs Fellow at the Carnegie Endowment for International Peace. Scott, we appreciate your time. Thank you. Thanks so much, Ed.

16:29After the break, a deeper look at AI debt. And for even more markets insights, you can subscribe to my weekly newsletter, Simply Put, at simplyput.profgmedia.com.

16:52and downs, peaks and valleys. Weight loss can feel the same way. Often losing the weight is only half the battle. Keeping it off is the other half. That's why Noom combines GLP-1 medication and behavioral support designed to help people build healthier routines and sustain progress over time. Noom found that 37 % more weight is lost when using Noom than just taking GLPs on their own. And the weight is sustained too. That's because Noom uses a psychology-based science-backed program that pairs your medication with daily lessons and food logging to support weight loss that lasts. Plus, you can get personalized weight loss plans and expert coaching designed to maximize your success with GLP-1s while helping you develop sustainable habits to live better and longer.

17:36The Noom GLP-1 program starts at$39 and is delivered to your door in as little as seven days. Start your GLP-1 journey today at Noom.com. That's N-O-O-M.com. Noom, Lose for keep. Initial one-month subscription from$39 plus tax and$99 plus tax per month thereafter. Cost of medication not included. Final pricing depends upon program selection.

18:04Support for the show comes from SoFi. Your home might be one of the biggest assets you own, but a lot of people don't realize you can actually put that equity to work without refinancing your existing mortgage. That's where SoFi Home Equity can help. Whether you're planning a renovation, consolidating high interest debt or covering a major expense, SoFi offers home equity solutions designed to help you access your home's value while keeping your current mortgage interest rate exactly the same. If you want a lump sum with predictable monthly payments, a SoFi home equity loan lets you borrow up to$750 ,000 or 85 % of your home's equity with a fixed interest rate.

18:39Or if you prefer more flexibility, a SoFi HELOC lets you access funds as you need them during a three-year draw period, so only borrow what you need when you need it. If you've built equity in your home, it can be one of your most valuable financial tools. Head to SoFi.com slash ProfG Home to explore your options and check your rate with no impact to your credit score. That's SoFi.com slash ProfG Home. SoFi Bank N.A. NMLS 696891. Member FDIC. Offers loans directly or through Spring EQ, a state licensed lender. NMLS 1464945. Terms and conditions apply. SoFi receives a fee for loans procured to Spring EQ.

19:22Running a business shouldn't feel like surviving a software group project. One app for accounting, another for inventory, another for sales, and somehow none of them talk to each other. That's where Odo comes in, an all-in-one business management software that brings every part of your business together. From sales and accounting to inventory and marketing, all in one powerful platform. No messy integrations, no bouncing between tabs, and best of all, no spreadsheets. Stop managing software and start managing your business with one unified system. Try for free today at odoo.com slash vox. That's odoo.com slash vox.

20:01We're back with ProfG Markets. The world's most valuable chip maker is making its lenders nervous. The cost of ensuring NVIDIA's debt against default just posted its biggest one-day jump on record. The spike followed reports that NVIDIA has lined up more than$750 billion in new AI commitments. That includes a$250 billion guarantee for OpenAI and a$500 billion deal with SK Hynix, his parent company. These developments have revived a fear that we've been discussing on the show for a long time, and that is circular financing. The worry is that deals like these create a loop. NVIDIA funds its own customers who spend that money on NVIDIA chips, making demand look bigger than it is, and investors are beginning to worry.

20:45The stock fell 5 % on the news, and it's now down over 16 % from its May peak. But NVIDIA isn't the only company under scrutiny when it comes to debt. Oracle's credit rating was downgraded this month to one notch above junk, and its default insurance is now priced at levels last seen in 2008, which begs a question that the market is struggling to answer. How real is the demand for AI infrastructure and how risky is the debt that is financing it? Here to help us discuss this, we're speaking with Vishy Tirupatur, Chief Fixed Income Strategist at Morgan Stanley. Vishy, thank you so much for joining us on the show.

21:25AI debt is suddenly the rage right now and it's something we have been discussing at length on this show. It's something that a lot of people seem to be getting a lot more nervous about. We obviously had Oracle and their debt situation, which has gotten trickier, and we've seen that reflected in the bond markets. Now it appears to be affecting NVIDIA. What do you make of all of this? What does it say about the AI credit markets at large?

21:50Vishy Tirupattur:I think what's happening is that a new sector is coming to the credit markets. If you look back just a few months ago, the AI ecosystem represented in the benchmark indices were in the investment grade market, you know, were under 3%. Today, they are over 6%, more than doubled in a very short period of time. So when you have this much of supply and the expectation that there is a lot more supply ahead and is the crux of the agita that is in the market, if you step one more, one step back and examine it, it is really the expectations of the CapEx from the AI ecosystem, the hyperscalers included, that is a pretty substantial amount of upward revisions of CapEx.

22:42Vishy Tirupattur:So beginning of the year, we thought that the CapEx by the five hyperscalers in this year would be something around$600 billion in 2026. and we thought the number would be around$850 billion or so for next year. And we now think that the this year number, pretty 26 capex of these five hyperscalers, would be greater than$800 billion. And the next year, we're talking about$1.3 trillion and similar magnitude in 28. So this is a substantial reset of expectations of capex. And a good chunk of this capex has to get funded through the trade markets. So if you think about it from that perspective, the expectations are constantly being revised as to incremental amount of supply that needs to be absorbed.

23:36Vishy Tirupattur:So keep in mind that, as you said, Oracle is at the low end of the credit sector, but the rest of the hyperscalers are very high quality. I mean, high quality Microsoft is AAA, Alphabet is AA, and Amazon and Meta or AA, A+. So much, much higher quality. But the expectation of a lot more supply in multiple currencies, in multiple maturities, in multiple forms, in unsecured form, securitized form, public markets, private markets, all of this abundance of issuance expectations that are ahead is the source of this agital. And if each time we have this expectation reset, then there is incrementally the spreads getting wider.

24:22Vishy Tirupattur:So in our mind, the credit spreads need to be wider to accommodate all of decisions. So not everyone is going to be equally wide. The riskier names, so for example, credit risk lower the credit rating and the greater the spread widening that you've seen. So as you mentioned, unsurprisingly, a name like Oracle is trading at a, you know, as of Friday, they were trading, you know, well over 220 basis points in spread terms for their benchmark ones. That was, that's higher from a year to date, almost 65 basis points higher. But it's not uniformly like that. So the more trade worthy names are wider, but wider by less.

25:12Something we read recently from Nikkei Asia was this report that, and as you mentioned, these hyperscalers, a lot of these names, their credit ratings are in pretty good standing. And then when you look at their balance sheets, they're in pretty decent financial health. But we read this report from Nikkei Asia, which reported that Meta, Oracle, Alphabet, Amazon, and Microsoft, all the hyperscalers have roughly$1.7 trillion in debt that is off of their balance sheets that has been registered through these SPVs that get funded or that get financed by these private credit funds. And so a lot of the debt that is financing the AI boom, we don't really see much of because it's in these SPVs and they're not really taking much of a role in the larger names of the bond markets.

26:02Does that worry you? If so, why? And if not, why not? And how much attention should we be paying to that number, that$1.7 trillion number?

26:12Vishy Tirupattur:From an investor perspective, you have to worry about all of the nuances associated with it. I think the traditional distinctions we've had between public and private, secured, unsecured, investment rate, high yield, secured, structured, all of these differences are sort of merging. These silos are merging. And we have bonds that are hard to fit into any one bucket. And more and more investors are getting used to thinking about these bonds from not just from one perspective, from a variety of perspectives. So almost all of these bonds are predominantly in institutional investor hands. and the distinction, each of these different forms have their own specific risk and return issues of conservations associated.