# What If China Wins The AI Race? **Channel:** Andrei Jikh **Source:** https://www.youtube.com/watch?v=a7eW1OU1v00 **Transcript page:** https://www.withtranscript.ai/video/a7eW1OU1v00 ## Chapters - 0:00 — China, Nvidia, and the AI Stakes - 5:59 — Open Source Battle and Market Pressure - 10:29 — Why the Story Surfaced Now - 14:38 — What Happens If China Wins - 20:41 — Investing for Both AI Outcomes ## Transcript **[0:00] Speaker A:** All right, so what happens if China wins the AI race? **[0:04] Speaker B:** Nvidia CEO Jensen Wong under severe scrutiny amid revelations that he had an affair with a suspected Chinese spy. Nvidia's board has launched an investigation. **[0:14] Speaker A:** Constantly throughout the tape, we see her grooming herself, fixing her hair. She doesn't want to look bad in front of him physically. And that's very telling. **[0:23] Speaker B:** This comes as concerns grow over Wong's interactions with Mega Speed AI executive Alice Huang, an alleged Chinese intelligence asset. **[0:32] Speaker A:** So that is the story. Now on the surface, it looks like a classic Chinese spy movie story. Right, China cannot build the best chips. It's in a race with the United States. So it hires a spy to get close to the man who does. And that is one way that China could win the AI race. But the more I looked into this story, the more I realized that this goes so much deeper than that. This is a story about money and control over how companies monetize AI and what the US stock market should even be worth. Should it be worth more than half of the world's stock market combined? Now before we get into who's fighting and why, let me explain what this fight is actually about. See, when most people use AI today, they use something like ChatGPT. They ask a question, they get an answer, but behind the scenes, what you're actually doing is you're renting access to a model that lives on some company's computer. **[1:32] Speaker A:** And that company decides what that model can say, what it costs, and whether you can keep using it or not. If they raise the price or they shut off your access, there is nothing you can do about it. That's called a closed source business model. Now on the opposite side of that are open source models. Open source is where a company publishes the model and then anyone can download it. You could run it on your computer, you can run on your servers, right? You can change the model however you want, and no one can turn it off or charge you money for it. Now why so important? It's important because a business that runs on closed source will treat AI like a product, something that the companies could own and sell to you. But open source, they treat AI like a utility, which is something everyone should have access to, right? **[2:25] Speaker A:** Like electricity or the internet. And depending on how you make money, you desperately want one or the other. That's what this fight is about. Now the question is, who wants AI to stay expensive and who wants it to be free and why? On the closed source side, you have the companies that sell the models. These are companies like OpenAI, Anthropic, Google. They keep their best systems behind paid access. For them, a free model that's almost as good, that's a huge threat to their business because why would you pay for something if you could get it for free, right? Now on that same side, you also have national security interests and you have the capital aspect of it. That's the money part of this story. And that's because hundreds of billions of dollars have been borrowed and spent building AI. And there is an assumption that someday AI is going to make trillions of dollars. **[3:21] Speaker A:** And this is why the US stock market is as expensive as it is today. But it can only pay it back to shareholders if those companies can charge you money for it. So that is one side of this fight. Now on the opposite side, we have the open source strategy, which are those who want it to be free. And we have a very different set of players here because NVIDIA, for example, does not make money selling AI models. It makes money selling the chips that run them. So the more AI there is in the world and the cheaper and more widespread it gets, the more chips NVIDIA can sell. That's why NVIDIA plans to spend something like $26 billion over five years building open models and giving them away for free. It's the same logic for those cloud companies. Free models mean more people running AI, which means more demand for what they sell. **[4:16] Speaker A:** So those are the two sides in what they want. Okay. Well, then there's China. China wants a very specific outcome cuz China cannot build the best chips yet. It's still behind on the most advanced hardware. But China also knows instead of trying to beat the US at selling AI, all China has to do is give its AI away for free, which is also why Chinese models are now the most downloaded open models in the world. It's also why most of the developing world is adopting them as a way to own their AI instead of renting it and paying a US company. So what does China want exactly? What it wants is for AI to be worth nothing because America has bet everything on it. Because if the best AI in the world is free, then the company is trying to sell it can't charge you money for it. **[5:09] Speaker A:** And that has huge implications for the trillions of dollars that was spent building it. Now, Jensen Huang is in a very strange position. He's an American CEO whose business model lines up more with China strategy, which again is to give it away. And that's exactly why, depending on which side you're on, aka how you make money and how you use AI, he's either the hero of this story or he's a problem and he needs to be dealt with. So in this video, I'm going to try to explain this and give you some evidence for all of this because ultimately I want to help explain if China wins this AI race, what happens to the trillions of dollars that the US has bet on AI? What does that mean for our money and our investments? So with that said, let's get into it. **[5:59] Speaker A:** Hi, my name is Andrej Jik. Hope you're doing well. Come for the finance and stay for the AI race with China. So first, let me show you some evidence for what this fight looks 想像 so far, because this summer, a Chinese company called Moonshot AI released a model called Kimi K 3. It was a 2.8 trillion parameter model, and it made a lot of chip stocks nervous because it beat one of the best US models on some of the benchmarks. And it was open source. It was free for anyone in the world to download. And that set off alarm bells for the US closed source way of business. Even Treasury Secretary Scott Bessent went on TV and said that the administration would examine Chinese models for intellectual property theft, and they could sanction the companies behind them. **[6:45] Speaker B:** I think that they didn't realize how powerful their open source models are. So, you know, the the open source models, they do industrial distillation, which is a nice word for stealing from from the US model. **[7:01] Speaker A:** And that is why, according to the New York Times, the administration was seriously considering sanctions and trade bans on China's open models. The companies pushing the hardest for these restrictions, by the way, were closed source labs like OpenAI and Anthropic. And why they're so supportive of this under the guise of national security is of course because if you ban the free Chinese competition, then the paid US models get to keep charging money for their AI, obviously, right? National security and business interests are on the same page on this. But then the other side of the model fought back. On July 24th, more than 25 companies, including Nvidia, Microsoft, and Meta, signed a letter telling Washington not to ban open models. And Jensen Huang, who had never posted on Twitter before, made that letter his very first post. He wrote that open models, quote, enable sovereignty. **[7:58] Speaker A:** What is sovereignty? Sovereignty for a person is control over their own life. And then days earlier, he also said American companies should be allowed to use Chinese models. Elon retweeted it. He said quote, this has my full support. Jensen is right. The administration backed off and they were like, okay, fine. We'll let the open source people do their thing, at least for now. So Jensen Huang has sort of become the face of the resistance and that was July. Then three months later, on October 29th, the New York Times published a report that US intelligence believed a woman close to him was passing information to China. Three Republican congressmen demanded answers from Nvidia. And obviously just recently Jensen Huang admitted to his affair. Then Nvidia opened that investigation into him. So basically the guy who stopped the US from banning the Chinese AI models is also now being investigated over his ties to China. **[8:55] Speaker A:** That's what's happening. Now is that a coincidence? The timing of it is very interesting and it made me want to look deeper at this and so I did and here's what I found. Now before I explain that, some of the best ideas that I come up with come from conversations that I forget to record. I spend a big part of my week researching upcoming videos and the most important thing usually gets said in the last few minutes when I've stopped taking notes. And for years I used my phone for this, but it drains the battery, sometimes notifications interrupt the recording, and then I have a 90-minute file that I'm never going to listen all the way through to find what I need. So now I use Plaud Note Pro. Plaud is basically my AI note taker. It's about the size of a credit card. It lives inside my wallet and it can snap magnetically onto the back of my phone. **[9:36] Speaker A:** And the way it works is I hit record, it transcribes the conversation, summarizes the key points into action items, and later I can ask Plaud something like, what did he say about oil? And it finds it. A recent example is when I was watching an FOMC video about interest rates for one of my videos. And since those sometimes can be very long, it comes in handy to recall ideas I think are useful in my research. It also works for in-person conversations, phone calls, and online meetings. Plaud is built around enterprise-level privacy and security standards and for me, a lot of what ends up in my videos starts as a If yours does too, whether that's clients, interviews, or meetings, this device will pay for itself. I spend less time taking notes and I spend more time being creative. You can use my code Andre 22 at the link in the description to get 22% off. **[10:23] Speaker A:** There's also a 30-day free return policy, so it's risk-free. Thank you Plaid for sponsoring this segment, and now let's get back to it. **[10:29] Speaker A:** Okay, so here's one of the most interesting parts to this China story that you might not have known about. According to the New York Times, Trump actually knew about Alice Huang roughly a year ago. And a separate Times report said that Marco Rubio allegedly oversaw a memo about the relationship back in 2025. But nothing happened. Jensen Huang still kept his close relationship with the president. Nvidia still kept selling chips, and businesses linked to Alice Huang kept on buying Nvidia chips well into 2026. And then in July, Jensen becomes the face of the open source resistance. Right, he sends out the tweet that stops the US from banning Chinese AI models. And then a few months later, the story finally goes public. So the question is not even about whether the spy story is real or not. It might be, it might not be. The real question is, why did this story become public now? **[11:26] Speaker A:** One theory says it's because the information has become useful. Right, so think about it from the opposing side. The side that's foreclosed source models. For a year, this intelligence was just sitting there. Jensen Huang was not a problem. But the moment he became the face of the fight against banning China's AI and then was actually winning, he also became the biggest obstacle for anyone who was making money from AI staying closed source. The easiest way to beat someone arguing that we should trust China's AI is to show that he himself is compromised by China. Now, we don't actually know who leaked this story or or But what makes this story so interesting is because if the closed source side wins the fight, it doesn't actually stop AI from being free. It only stops American AI from being free. And people don't want to pay for AI, which is why China's models now make up the biggest share of downloads. **[12:32] Speaker A:** Most of the developing world sees China as the best path to owning their AI. The biggest thing stopping the closed source model of business was Nvidia trying to be a provider of open source models. But if Jensen Huang gets pushed out and replaced or he magically changes his mind about supporting open source for some reason, right? Then that means America's open source movement is done. Our corporations move to a fully closed source system. And the ironic part about that is that the people trying to stop China from winning might actually end up giving China an easy win. Because last week, Democrats swept Congress and they're already preparing new AI rules, including mandatory safety testing before models are released and new liability for companies whose AI causes harm. All of that sounds like the government regulating big tech, right? But what that's actually going to do is it's going to give them the moat they always wanted, the protections, right? **[13:41] Speaker A:** Because who could afford to follow these rules? Only a giant closed source lab which has an army of lawyers and a compliance department. Because an open source project that gives away its models for free and is not making money has no resources to defend itself against big government coming after them. So even though these rules sound like they're going to rain in the big AI companies, what they'll actually end up becoming is just rules that will end up protecting them, which is what those companies have always wanted. This theory says that that's exactly what's going to happen. And if that happens, the American open source gets regulated out, the world will still keep on using open source. It'll just be all of the Chinese models, and China wins anyway. The irony is that the short-term thinking of our US politicians and their greed is going to cost them the race in the long run. **[14:38] Speaker A:** And that brings us to the real question of this video, which is, okay, what actually happens to the world and to the US stock market if China does win this race? Okay, so what actually happens if China wins? I don't think anyone really knows for sure, but there's a really good theory that helped me to understand it. So credit to Luke Roman from FFTT. The link to his research is down below in the description. But he basically says that there's two ways that this will go. Either AI works or it doesn't. But either way, the scary part is that both outcomes lead our economy to the same place. If AI does not work, the trillions of dollars the corporations borrowed to build it blow up. That's not good. If AI does work, then it replaces the jobs the government depends on for taxes. That's not good. **[15:25] Speaker A:** Either way, the government's only way out is to print more money. And printing money leads to an inflationary collapse, meaning everything goes up in dollar terms but goes down in real money terms. And if that sounds confusing, let me explain it. So outcome number one, let's assume AI does not pay off. What happens then? First, you have to understand just how much money is being bet on AI working. That's why I suspect that we're seeing such a big push for closed source models right now. That's why we're seeing anyone who might be in the way of that model being handled, allegedly. Right? These companies and their profit incentives are trying really hard to keep AI expensive so that someday they could pay back their shareholders what they spent building it. But there's a problem. One executive at one of these AI companies told Lou Grohman that the top end model he had in his computer four months ago is now free. **[16:24] Speaker A:** What that means is if the best model in the world goes from premium to free in just four months, then the company that built it technically has four months to make all that money back. Otherwise, that money's gone. Right? Luke says that the capital's already been burned and it just hasn't been marked to market yet. And that's before you factor in China. Chinese open models are now roughly 90 to 95% as good for about 1/100th of the price. And China might want this to happen because it puts pressure on the whole US system. Now it actually gets worse from here because these AI companies are not just spending their own money anymore. They've started borrowing huge amounts of it. And all that borrowing competes with the US government for the same pool of money. And ultimately, that pushes interest rates higher. What do I mean by this? **[17:16] Speaker A:** So think of it like this. There's only so much money in the world that's available to lend. Banks, insurance companies, investors, they all have a limited amount of money that they're willing to lend out. The US government borrows from that pile by selling Treasury bonds. Right? Now AI companies are borrowing from the same pile by selling their own bonds. So the lenders have a choice, right? Lend to the government or lend to the AI companies. When there's more competition for the same pool of money, the price of borrowing goes up. The government is trying to incentivize investors to lend them money instead of the AI companies. And the way they incentivize is obviously by offering higher interest rates to convince people to keep buying its debt. The AI companies can afford to pay those higher rates because they think their payoff is going to be huge. **[18:11] Speaker A:** But for the government, that means its interest bill on the trillions of dollars debt just keeps growing and growing. That puts pressure on the US. And because Treasury rates set the baseline for almost everything else, when they go up, mortgages, car loans, credit cards, aka our lives become more expensive too. So if AI can't pay back that debt, then you've got hundreds of billions of dollars that need to be written off and a stock market that's built on a handful of tech companies that suddenly aren't worth what everyone's worth, what do you get? Boom, right? You get a deflationary collapse. That outcome needs a bailout. Where does the government get the money to do that? It has to create it, right? That leads to an inflationary collapse. So that is outcome number one. Outcome number two is AI works, right? We get all the benefits, max productivity. **[19:11] Speaker A:** But with productivity, we need less jobs, right? More efficiency. Luke Gromen says that this is just as dangerous because the government is now competing for money against the very industry that's building a system to replace the workers it needs because it needs them to collect taxes. In other words, America's borrowing against its own future to fund a technology that is designed, at least in part, to take away people's ability to make money, to borrow money, and to spend money. Okay, how's it doing that? Well, if AI succeeds, it threatens the source of about half of all US tax revenue, which comes from people's jobs. So if AI works the way those companies promise and it replaces white-collar workers, then tax revenues fall while the government's bills keep growing. Where does the government get the extra money from? It has to create it, right? Either way, they have to create more money, so the government is stuck. **[20:14] Speaker A:** It can't let AI fail because too much of the economy now depends on it. There's too much financial incentive for them to just let it fail. And it can't let succeed either, without losing its tax base. So, either way, it eventually has to rescue it, no matter the outcome. And the only way to do that is to create more money. All of this essentially forces America to choose between saving the dollar and saving AI. **[20:41] Speaker A:** And Luke Gromen thinks that they'll choose to save AI. Now, how does China play into all of this? China doesn't really need to build the best technology. Maybe they steal it from time to time. But all China has to do is keep giving away their AI that's almost as good for almost nothing. And then they wait for the math to break the world. What's kind of interesting to me is that since the launch of ChatGPT in November 2022, gold has only underperformed the Nasdaq by about 15%. I guess that's not a lot for something that people called a shiny pet rock that's not productive in any way. Ultimately then, what does this mean when the most revolutionary technology of our lifetime is only doing just a little bit better than a pet rock? What does that mean for our money? Right, here's what I think. **[21:35] Speaker A:** Okay, so if you made it this far, this is obviously not financial advice, but if Luke Gromen happens to be right and both outcomes lead to the money printer, then the question for me is not really whether AI wins or loses. That's a coin flip at this point. I don't know the outcome. The question for me is what do I want to own when the printing starts? And the way I think about it is the same two outcomes that we talked about. Right, in outcome number one, AI doesn't pay off. The debt blows up, private credit markets blow up, life insurance blows up like we talked about in another video. And we get this deflationary collapse first. In that world, I think cash and gold go down the least because they've been the thing people run to when they're scared and they don't trust the system. **[22:20] Speaker A:** Now, in outcome number two, let's say AI does work, so the tax base gets smaller, and the government prints its way out. It saves the system, right? That is more of an inflationary collapse, a collapse up where everything goes up in dollar terms. And in that world, I think Bitcoin does the best. Luke Gromen also pointed out that the last time we did this during 2020 and 2021, gold took off, but Bitcoin really took off. So, I don't personally have a view on which outcome happens because I don't know. Which is why I want to own something for both outcomes. And it's also why I recently started buying physical gold and physical silver, why I hold my Bitcoin in self-custody, because I think gold protects me if things collapse downward, but Bitcoin protects me if things collapse upward. Luke Gromen also pointed out that the US has not added to its electricity generation for 20 years, from 2004 to 2024. **[23:21] Speaker A:** So, whether AI succeeds or it doesn't and it's all a bubble, the country still needs a lot more power than it has. It needs a lot more grid and a lot more copper and silver to build out that future. Which also means it's probably smart to own things that are hard to print, like energy, metals, and like hard money. At least that's my understanding of it. I'm curious what you think. If you want to know more of my thoughts and how I'm thinking about investing, all of those videos and more can be found in the premium member section where I post early and sometimes I post my extra thoughts. It allows me to make more videos like this one and take on fewer sponsors. Thank you so much for watching. I'd love to see you here next time. Take care.