Transcription
And I predict we're going to have 5, 10, 15 deep seats in different sectors in the next generation. The auto industry is a pillar industry of the world. It took three countries 75 years to dominate that pillar industry. Okay. Germany, Japan, and the US. And in 5 years, we uprooted the entire thing. Yeah. And it's never going back. Hands down.
Welcome back, guys. Here's the reality. China is going to dominate every major tech sector in the next decade. Not because they're stealing, not because of some temporary advantage, but because they've built a system that systematically outcompetes the West. And once you understand how they're doing it, you'll see why it's unstoppable. Let's get into it.
Welcome back. A new report claiming that despite export bans, Deepseek used NVIDIA Blackwell chips to train its latest AI model and is expected to drop next week. And that's not the only new accusation being leveled against Chinese AI labs. Dear Jabraosa has more in today's tech check. Hi Dearra.
Hey Kelly. So the argument for selling chips into China was always very simple. If we sell them older generations of American chips, they'll buy from us instead of building their own and having a market for their own homegrown chips. Now Jensen Huang has been the loudest voice making that case along with David Sachs and others in the Trump administration. But what we are learning is that China is not choosing between buying and building. They are doing both. It's a very large market. Plus, they're stealing the software through as well through distillation, which is running millions of queries against American AI models to essentially extract the intelligence inside of them and replicate it. Anthropic is the one making this claim.
Okay, so right here, stealing the software through distillation. This is the core narrative they're pushing. But here's what they're not telling you. Deepseek made about 150,000 queries to Anthropic's model. That's it. 150,000. Do you know how much traffic a small AI app gets in a day? A lot more than that. So this number is a rounding error. This is not evidence for systemic theft. This is Anthropic trying to create a regulatory capture narrative so the government will ban Chinese AI to protect their market share. I mean, I guess you can say it's a brilliant strategy that has worked for other companies, but it's not based on real evidence. If you want a way more detailed breakdown of why Anthropic's claims are fake, I'll leave a link to a video that goes deep into this. But the short version is 150,000 queries is nothing. It's just benchmarking. Every AI company does it.
Now, so the two things that the US was counting on to stay ahead, that is controlling the hardware and protecting the models, those are both failing. And that is tough for everyone who argued that export controls were the wrong approach. It also raises a harder question. If China can get the chips anyway, clone the models, where is America's moat? Now, look at this. Four of the five most used models in the world are Chinese, according to Open Router. That doesn't mean that they're the best, but they are the most used because good enough often outweighs best-in-class, especially when it's cheaper and open source.
They're not the best. That's not even true anymore. Deepseek 2.0 is competitive with GPT4 on most benchmarks. But here's the thing that really matters. They're selling a million tokens for $0.38. OpenAI is selling them for $5.50. That's a 14-fold price difference. So, you've got best-in-class quality at a fraction of the price. Of course, they're going to be the most used. That's the inevitable result of the West trying to sell luxury products while China is turning technology into a utility. And here's something wild. When you ask Claude 4.6, Anthropic's own AI model, what AI model it is, it confidently answers, "I am Deepseek." Anthropic is accusing China of stealing their technology while their own model is basically saying it's been replaced. That tells you everything you need to know about what's actually happening.
Now, Kelly, this all makes for a very interesting setup for Nvidia earnings tomorrow. The case for keeping the China market open is getting a lot harder to make and as you mentioned at the top, we are anticipating that next Deepseek model sort of any day now.
Yeah, you know, I will say again, credit to Jeff Curry for making me think a little bit more about this. Dearra, the China AI threat, it seemed like it was kind of on the back burner, but it becomes relevant. And maybe we can put up that chart you showed again of the most used AI models are often those that are the cheapest and easiest to access. At some point, if users perceive this as somewhat of a commodity and if China makes, you know, has a cheaper production method for making, then they could do to AI and to our Mag 7 the same thing that they did with global exports.
This is actually the most honest thing that's said in this whole segment. Yes, China is going to do to AI exactly what it did to manufacturing. They're going to dominate it. They're going to make it cheaper. They're going to scale it globally. And the West won't be able to compete on price. And the fact that CNBC is presenting this as somewhat of a bad thing shows you they don't understand what's actually happening. You see, here's what they're really missing. This is good for the world. When China commoditizes AI, it doesn't just hurt American tech companies, it liberates the global south. A startup in Nigeria, Brazil, or Indonesia can now build AI-powered products for $0.38 per million tokens instead of $5.50. That's the difference between being able to afford cutting-edge technology and being locked out forever. Think about it. A developer in Lagos can now build the same AI application as a developer in Silicon Valley at a fraction of the cost. That's democratization. That's the global south finally getting access to the tools they need to compete globally. And that's why China is winning, not just in tech, but in soft power, in influence, in shaping the future. The West is trying to protect their market. China is expanding the market and winning the customers.
With on so many fronts. I feel like, you know, this is something that our US markets have become so complacent with, treating Deepseek last year sort of as a one-off and as a blip. And I've been arguing the opposite for it feels like over a year now. It's not that Deepseek was a blip at all. The adoption rate has really been amazing. I mean, not just in the US. They've actually gained traction here, but all over the world. And you're right, as models become more of a commodity, as distillation is a technique that the Chinese can use at a more efficient rate, it's these Chinese open-source models that have been gaining traction globally. They're not frontier models.
Of this top 10. So what's Minimax? What's Qwen? What's GLM? What's Trinity? I think I know the rest. Grok and Claude.
Yes. Okay. So these are all sort of the top Chinese AI labs backed by Alibaba. GLM5 is a company called Zhipu AI, which just went public in Hong Kong. I believe Qwen is Moonshot. Um, it's amazing because, yeah, a lot of Americans have not heard of these models or these labs. Um, but as you see here, they're the most used. So they're not the best models, but they're the most used. And when you have sort of these viral things like AutoGPT, right, agents that require so much compute power, a lot of folks are switching to these cheaper, lighter models that are good enough because you're just using way more.
Not frontier models. Okay, so what? Frontier models don't win markets. Better business models do. Look at Netflix versus Blockbuster. Netflix's technology was better. Streaming beats DVDs. But that's not the only reason it won. It won because its business model was better. No late fees, no physical stores, simple subscription. Blockbuster was extracting money from late fees, $800 million a year. While Netflix was building customer loyalty. Same thing here. Chinese AI's business model is better: open-source, affordable, accessible. The US model is: build it, charge a premium, protect your market. China's model is: build it, make it cheap, dominate the market. One of these business models wins. It's not the US.
Now, let's talk about the real story. This is Eric Lee, a venture capitalist and political scientist who actually understands China's long-term strategy better than almost anyone. What he's about to explain is why China's dominance isn't a surprise.
The the 15-year plans.
Yeah. I mean, you know, Mao had already this vision that, you know, within 75 years, China could overtake the US, right? I mean, he he said that at the at the announcement of the of the first five-year plan, and it was crazy. It was um, he, we're now starting the 15th, correct, five-year plan. And he wrote something to the effect that, hey, you know, we're starting our first five-year plan. Let's work hard and and look, our goal is not to mess around and just just put food on the table. Our goal is to catch up and surpass at the time the most advanced country in the world, the United States. He says, "How many years is it going to take? 5, 10, 20? Probably too soon? That's too optimistic." He said, "I don't Oh, he said maybe maybe 75 years. How about that? That's just 15 five-year plans." He actually wrote that at that time. It was 1950. He was pretty prescient. Let's put this into perspective. In 1950, China was one of the poorest countries on Earth. Per capita income was around $89. Countries like Congo, Ivory Coast, Ghana, Kenya, Nigeria, and South Africa were all richer than China. China's per capita income was only one-third of Sub-Saharan Africa's. They were that poor. Fast forward to 2025, China's GDP is now around $45 trillion, the largest economy in the world in terms of purchasing power parity. Per capita income is around $24,800 in purchasing power. They've lifted over 800 million people out of poverty. Their poverty rate fell from 88% in 1981, not too long ago, to 0.7% in 2015. They now control 35% of the world's industrial capacity, more than the US, Germany, Japan, India, and the next five countries combined. This is the most rapid economic transformation in human history. No country has ever gone from being one of the poorest to having the largest economy in a single lifetime.
You know, I mean, you're you're up there now, you know, when it comes to a bunch of stuff.
EV for sure.
We're now.
AI pretty soon.
We're embarking on the 15th five-year plan and I'll tell you what's happening in China. Okay. Uh, amazing stuff. Okay. We are so so we began to engage in globalization 30 some years ago, right? And it of course began to uh, it it we joined the WTO in in 2000. Um, and the way we engaged globalization at the time was through manufacturing, uh, because that's our strength at the time. And we went from single-digit share of the global industrial capacity to today, our industrial capacity is the biggest in the world, bigger than the US, Germany, Japan, India, and the next five, six countries put together. This is the foundation. This is what allows them to scale. You can't dominate AI without the ability to manufacture chips at scale. You can't dominate biotech without the ability to run clinical trials at scale. You can't dominate robotics without the ability to produce them at scale. China has all of that. The West doesn't. And now they're taking that same manufacturing expertise and applying it to high-tech sectors. Okay. Enormous. 35% of the industrial output, I think going up maybe to 40%.
Yeah. Okay. Um, and so so we kind of cleaned up industrial capacity on manufacturing and very successful. Okay. Um, but at that time, 2000 plus minus, uh, we weren't in the room in terms of science and technology.
Right.
Okay. And the US was indisputably the leader. So we just followed Silicon Valley. Okay. Whatever they did, we we we followed. And we seeded enormous territories like semiconductor, okay? Um, and and they were leading the Okay. But we didn't just lay flat like the you the the current popular Chinese term is, you know, "ting." We didn't lay flat. They they worked very hard on education, on science and technology infrastructure, all that. So it accumulated over decades. And we are now at the cusp of major science and technology breakouts and breakthroughs across multiple sectors. And I predict we're going to have 5, 10, 15 Deepseeks in different sectors in the next generation.
10 years.
Nonlinear.
Nonlinear, nonlinear. Okay. I'll just give you an example. Okay. I I I would say that this process began in 2020. Okay. And the first sector that it affected was renewable energy.
Okay.
And in 2020 to 2025, today, in those five years, not just solar and wind and all of that, the Chinese were very successful globally also in automobiles. And and let me kind of explain how how big that is. Okay. The auto industry is a pillar industry of the world. Okay. A pillar industry. Okay. And post World War II, it took three countries.
Yep.
75 years to dominate that pillar industry. Okay. Germany, Japan, and the US. And in five years, we uprooted the entire thing.
Yeah.
And it's never going back.
Hands down.
Five years. Think about that. Germany, Japan, and the US took 75 years to dominate the auto industry after World War II. China did it in five years. That's what you call execution. That's what happens when you have industrial capacity, capital, talent, and a clear strategy. And now they're going to apply the same playbook to AI, biotech, and robotics. The West has no answer for this.
Okay. 5 years compared with 75 years in a pillar industry in the world. Okay, that's auto. Okay, so from 20 right now, 2025 to 2030, in the next three year, uh, 5 years, I would say it's going to happen. Same thing is going to happen in three industries. Okay, biotech, which is already happening, AI, and robotics or advanced manufacturing, let's say, not just robotics alone. Okay. In biotech, and I began to invest in biotech 7, 8 years ago. Never in my wildest dream did I anticipate the current in situation. I mean, China is quickly becoming a biotech superpower. Okay. 7, 8 years ago, our share of the global novel medicine patents was maybe low teens, 11, 12%, 14%. Today is 44%. We'll go up to 50% very soon. Okay. Of all the clinical trials that are taking place globally, our share is 35% maybe, and and going up. That's bigger than the US, Europe put together. Here's the key takeaway in my opinion. It's not just about patents. It's about access. China is becoming the world's pharmacy. Companies from America, Europe, Russia, they're all going to China to source novel medicines. This is a complete inversion of the old model where the West developed drugs and the global south bought them. Now the global south is buying from China, and China is making them affordable. A drug that costs $10,000 in the US might cost only $1,000 in China. That's pretty much going to reshape the global healthcare system. But, you know, going back to price, let's be real about why the price difference is so massive. Big Pharma in the West is exploitative. Point blank. They charge Americans 5x, 10x, sometimes 20x more for the same drug than what patients pay for in China or Europe. It's not because the drugs cost more to make. It's because they can. They have patents. They have monopolies. They have a broken healthcare system that allows them to price gouge. China comes in with the same drug at a fraction of the price. And suddenly the global south realizes they've been getting robbed for decades. This is why Big Pharma is panicking. It's not because China is stealing. It's because China is exposing how exploitative the Western model is. And this is just the beginning.
Okay. The entire world is here in China shopping for novel medicines. I always said that, you know, in the year 2000, if you want to get a picture of globalization, you go to Guangzhou to the Canton Fair, right? Everybody in the world is in Canton Fair. If you own a little gift shop in Seattle or you own a uh, whatever, um, factory in in Barcelona, you're there buying whatever you need for your home markets. Today, there's a Canton Fair happening in China on biotech. Okay. All these companies are entertaining potential customers from America, Europe, Russia, and you name it, buying novel medicines, IPs from China. Okay. And that's happening already. So in the next 5 years, I think it will continue. Okay. Second, AI. AI, of course, everybody's saying US and China compete. Of course, there's competition, but China takes a totally different approach, and it's very relevant to the global south. Let me explain.
Um, so we take an open-source approach. Okay. US takes the closed-source approach. It also has to do with political system and political DNA. Okay. Will take hours to explain this, but let me just I'm with you.
So the US US approach is essentially invest, innovate, and see grants.
Correct. Okay. Like all these companies, all these big techs, we've been paying them rents for the last 20 years. You know that. Okay. Um, the Chinese approach is different. It's invest, innovate, and compete and scale up and make things affordable. That everybody gets it. Look at the numbers. In manufacturing, China produces goods at a fraction of the cost. An Apple Watch equivalent in Shenzhen might cost, you know, $15, $20, $30 bucks. The same thing in Palo Alto cost $450. That's a 30-plusfold price difference. Same quality, same functionality. That's the Chinese model in action. And they're applying that to every sector. AI, $0.38 per million tokens versus $5.50. Biotech drugs at 20% of US prices. This is about efficiency, scale, and a fundamentally different business philosophy.
Okay. Like like what we did with manufacturing. Okay. How come, you know, how come everybody in America, I mean, you think all these Americans and Europeans could afford to buy these Christmas toys 30 years ago without China? No.
Dude, you go to Amazon, you go to Walmart, 99% of the goods remain in China.
You can afford. Of course, that has an issue. There is an issue. I understand where, you know, I think President Trump said, you know, maybe too many dolls is not a good thing. Maybe two is enough. Maybe he's right. I'm not judging that. Okay. I'm just saying we make things massive scale and affordable. And same things happening in AI, okay? I mean, I know there the US is stronger on on computing power because the chips. Um, but it doesn't that's not where it's at, I think. Okay. We where I know companies around the world, including major American companies, are kind of secretly, they're not advertising, they're using Chinese AI. They're using Qwen, they're using, you know, Qwen, whatever it is. I mean, because look, the Chinese AI companies are selling a million tokens for $0.38 US.
Disproportionately cheaper.
Okay. GPT, I think sells them $5.5 US.
Yeah, American companies are using Chinese AI. I don't think that's such a secret. I mean, some are doing it secretly because it's politically sensitive, but others are doing it openly. For example, Airbnb CEO publicly stated they're running on Qwen models because they're faster and cheaper than OpenAI. Nvidia, Perplexity, and Stanford University are all using Qwen models in their work. And here's the thing. Roughly 80% of US AI startups that use open-source models are building on Chinese open models. 80%. So it's not some fringe thing. You can pretty much say that's mainstream. Why? Because $0.38 versus $5.50 is not a marginal difference. That's a huge price difference. You can't compete with that. And the fact that even major US companies like Airbnb and Nvidia, you know, the chip makers, are publicly choosing Chinese AI tells you everything you need to know about where the market is going. I mean, that's so big. The the difference is so big. Okay. I tell you, you know, without Chinese AI, global south and AI will never meet. They will have nothing to do with each other other than the global south paying rents to American AI companies, like they've been paying rents to Apple and Google and everyone else, all the big techs. Okay. But Chinese AI, open source, they are affordable. If you're an Indonesian company, you want to use AI, $0.38, maybe going down to $0.25 a million tokens. You can actually use them as a tool to help your business. It's actually affordable. You're not paying rents. Without Chinese AI, the global south has no choice but to be exploited by American companies. They can't afford GPT4. They can't afford the closed-source proprietary approach. But with Chinese AI at $0.38 per million tokens, now they can actually build. They can innovate. They're not just consumers anymore. They're producers. This is how you win the global south. Not through military might or political pressure, through affordability and access. China understands this. The West doesn't.
Okay. So, so I think, um, AI, very important, uh, especially in the global south. It may even be in America, but especially in global south. Um, and the third is, uh, advanced manufacturing, robotics. I mean, that's already happened. So that's the next 5 years, 2025 to 2030. And look beyond 2030 to 2035.
Oh my god.
I see future. You know, I can name quantum and nuclear fusion and I.
Well, you've already announced a thorium.
Yeah.
Right. Nuclear.
But still early. Okay. I see tremendous amount of capital, human resources, government policies going into these sectors. And I think will bear fruits 5 years from now. So we will be China will be at the forefront of science and technology across the board. Two two that will be made to be shared with the global south.
Think about that.
Two two fundamental observations that just keep on staring at me. First is the US economy is just too bloated. You know, you you take a Didi car ride in Shenzhen.
There's no tipping.
Right?
And on a per mile cost, it's cheaper. And the marginal productivity for China, I mean, you you can get an Apple Watch equivalent in Shenzhen for $12. It functions. It looks just as good when you're paying about $450 in Palo Alto for an Apple Watch. I mean, you know, for somebody in Africa, somebody in Papua, somebody in any village in the global south, those are just going to resonate, right? That's right. The way you open source everything. So, what could stop this? I mean, well, let me let me also share that you produce 4 to 4.5 million STEM products per year. Southeast Asia only 750,000, of which Indonesia 250,000. The US only 800,000 STEM products per year. Your your marginal productivity is only going to keep going up and beating the rest. But here's what really matters. It's not just the quantity, but what they're doing with that talent. China is taking millions of engineers and deploying them in strategic sectors. Biotech, AI, robotics, quantum computing, nuclear fusion. They're throwing human capital at these problems at scale the West can't match. The US has been outsourcing manufacturing for 30 years, which means they've been outsourcing the development of engineering talent. China has been doing the opposite. They've been building engineering capacity for three decades. Now they have a talent advantage that compounds every year. Unstoppable, right? So your your I think your struggle is the degree to which you can democratize capital to the global south.
Yeah. Would would you agree with that?
I I so so if you go to China, the buzzword today is called "Chuhai." Chuhai means literally means "going overseas." Okay. But but what it means are or Chinese companies, mostly technology companies, some manufacturing companies too, but but manufacturing companies with heavy technology, uh, content, uh, going to globalize, going overseas. Um, and a lot of these companies, including EV companies, I mean, I think BYD manufacturing in Brazil, in Thailand, in in in Hungary, Indonesia.
Indonesia. Chinese drone company, robotics companies.
Um, so all these comp battery companies.
CTL is opening a factory.
That's right. That's right. Um, so so I would say that look, I say I just throw this out for for I mean, I really haven't thought this through, so I could be wrong. Okay. Um, I I think the first phase of globalization, which is ending, has three main drivers. Okay. American capital and technology, Chinese production, and global market. Yeah. I think the next phase of globalization will also have three drivers: Chinese capital and technology, world production, and global market.
And and the Chinese have to spread their production and spread their their industrial and technological footprint and around the world, especially in the global south. And the the challenge is how to make it mutually beneficial. How to bring up the development of the entire global south, uh, uh, with that. And I think, um, the the political desire is there.
Yeah.
The geopolitical necessity is also there. Um, and the commercial incentives are there too.
So basically, the first phase of globalization was American capital and technology, Chinese production, and global market. That phase is ending. The next phase will be Chinese capital and technology, world production, and global market. So, China is going to be the innovation leader. China is going to be the capital provider, and production is going to happen everywhere: Brazil, Indonesia, Vietnam, Hungary, Africa. This is the new world order, and it's not coming in 20 years. It's happening now. If you found this breakdown valuable, if you appreciate data-driven analysis over mainstream media panic, hit that like button, subscribe to the channel, and turn on notifications so you don't miss the next one. And drop your thoughts in the comments. What do you think? Is China's dominance unstoppable? What sectors do you think they'll dominate next? Thanks for watching, guys, and until next time. Bye.