Speaker A
So the whole US stock market, including your 401k, your index funds, and the value of your retirement, is based on a story that might be coming to an end.
Explores how China’s AI strategy challenges the US-led AI market bubble and the risks corporations face with current AI investments.
Key Takeaways
- The AI market bubble is potentially larger and more impactful than the dot-com bubble.
- China’s cost-effective and open AI strategy threatens US dominance in technology.
- Current AI business models lack clear ROI and pose risks to corporate data security.
- Government restrictions on AI access are causing international partners to reconsider reliance on US tech.
- The bubble may burst when corporate spending on AI slows due to lack of measurable value.
Summary
- The US stock market’s AI-driven growth may be a bubble larger than the dot-com era, fueled by high expectations of perpetual American tech dominance.
- China is aggressively competing in AI by spending less and offering cheaper or free AI models, challenging US tech supremacy.
- US government restrictions on AI companies like Anthropic have led to international backlash and reduced trust in American AI providers.
- Many enterprises are frustrated by paying for AI token usage without clear ROI, as AI models often hallucinate and produce unreliable outputs.
- There is significant concern about data security and intellectual property risks, as AI vendors may learn from client data and become competitors.
- AI companies charge per token rather than on business outcomes, which raises questions about the true value and sustainability of AI services.
- The video includes insights from industry leaders like Palantir’s CEO Alex Karp, who highlights trust and pricing issues in AI adoption.
- China’s open-source approach to AI allows widespread access and innovation, contrasting with the US’s expensive and restricted AI ecosystem.
- The AI bubble may pop when corporations reduce capital expenditures on AI due to uncertain returns and rising risks.
- Financial indicators such as debt spreads and investment patterns suggest caution but do not yet signal an imminent collapse.
Chapters
- 00:00Introduction: The AI Bubble and Market Risks
- 01:46International Reactions to US AI Restrictions
- 03:22Interviews with Industry Leaders on AI Challenges
- 05:05Corporate Frustrations with AI Token Pricing
- 06:43Data Security and Intellectual Property Concerns
- 08:28Advantages of Running AI Locally and Open Source Models
- 10:06Media Coverage and Political Perspectives on AI
- 11:40Financial Implications and Signs of the Bubble Popping
Full Transcript — Download SRT & Markdown
Speaker A
You mentioned earlier the dot bubble. Are we doing bubble 2.0 right now? Oh, this is much bigger. The AI, the AI buildout relative to the TMT buildout of 1999-2000 is multiples, even as a percent of the economy.
Speaker A
Okay. So, one of the big reasons why the stock market is being held up right now is because there's a story that American companies are going to make trillions of dollars in profits forever because the world will be forced to use America's
Speaker A
technology. You have a lot of data that you look at. Do you think China is getting better at AI? How is the two there are two relevant tech centers on two and a half America, China, Israel. Those are the tech centers of
Speaker A
the world. They will win or we will win. Now, some people say that that story is coming to an end because of all the lies, the spending, and the competition.
Speaker A
Now, on June 12th, something interesting happened. A letter was sent to a company in San Francisco. That letter was from Howard Lutnik, the commerce secretary of the United States. And by the end of the night, the most advanced AI in the
Speaker A
world was shut down. Anthropic, which is the company behind Claude, was ordered to cut off two of its most powerful AI models from every foreign national in the world, not just China, by the way.
Speaker A
That order included countries like France, Germany, Japan, and even Anthropic's employees. If they weren't American citizens, they were also locked out. Now, four days after that letter, France fires Palantir. The French prime minister said that we can't depend on
Speaker A
partners who are capable of turning off the tap. But Germany already walked away. Spain told its companies to stop signing deals. The same goes for Britain. And it's because the world found out that it has a choice. And the
Speaker A
choice is literally seven and twelve times cheaper. Because while America is spending $1 trillion a year building AI, which is 3% of the whole US economy, China is spending a fraction of that and giving it away virtually for free. So in this
Speaker A
video, I want to explain how China is competing with the US and how this AI story might be coming to an end and some of the things that you can use to potentially see this bubble popping before anybody else. So with that said,
Speaker A
let's get into it. Hi, my name is Henri Jick. Hope you're doing well. Come for the finance and stay for the AI bubble everyone saw coming. So, let me just start with a basic question. Do people really want this AI technology? Because
Speaker A
there's a theory that says the reason that this is such a prevalent story in the market is so that these tech companies could justify their insanely high stock prices because in reality they've run out of really good ideas. In
Speaker A
fact, there's a really good interview on CNBC with Ed Zitron who brought up a lot of really great points.
Speaker A
But fundamentally, large language models are not the future. The only reason big tech is investing in this is that they've run out of hypergrowth ideas.
Speaker A
They don't have a next iPhone. They don't have a new Google search. So, they've put over a trillion dollars with trillions more to come into a kind of a dead-end industry because when that ends, they'll have to admit that
Speaker A
they don't have anything else. Now, throughout the video, I'm going to show you more clips from that interview, but there was also an interview with Alex Karp, who is the CEO of Palantir, which if you don't know is the company that
Speaker A
works closely with the government and pretty much every three-letter agency in the world. And Alex also brings up the fact that nobody really trusts AI right now.
Speaker A
Who owns the data? Where is it cached? Are the prompts secure? Is this being transferred to you? Are you being comp?
Speaker A
Okay, if it was so valuable, let's say I can make you a billion dollars right tomorrow. Wouldn't I say I'll make you a billion dollars and I want 30%? Why are they charging for tokens if it's so valuable? He is saying if the promise of
Speaker A
AI is as good as they are marketing it to be in its current form, they would not be charging us for tokens. Instead, they'd be charging us for building a billion-dollar business idea where they would take 30% of the revenue. Because think
Speaker A
about how you pay for anything in business. You pay a lawyer to win a court case. You pay a contractor to remodel your house. The price you pay is attached to a specific result. Now, AI companies do not work that way. They
Speaker A
charge us per what's called token usage. Now, a token is basically a word. Every word the AI reads and every word it writes for you, we pay for that, whether the answer was good or was really bad.
Speaker A
And Alex Karp is basically saying why would they price their business that way? Why wouldn't OpenAI ChatGPT just say only pay me when it works? If we create a good idea for you, give us a cut of your income. But I'm telling you
Speaker A
in this country at every single enterprise I deal with, these people are livid. They're like, I am paying for tokens that create no value. These people are stealing the weights and alpha of my business and they're creating a wealth tax that does not help
Speaker A
the poor. It just punishes starts with the billionaires. Every single person at this table is going to be paying a wealth tax only to punish us.
Speaker A
If they were confident that this thing created this value, that would be the easiest sales pitch in history. Pay us nothing unless we make you money and unless we build you a billion-dollar idea. But they can't offer that because
Speaker A
these models do what's called hallucinate. Right? This is where they confidently make things up. And nobody, including the people who built them, could tell you when or really why it happens. Right? No one's been able to figure out how to fix it completely.
Speaker A
You'll notice that both Anthropic CEO Darma Day and Sam Altman have both said, "We can't wait to see what you build with this." Well, that's because they don't know what you can build with this.
Speaker A
They want everyone else to do their innovation for them. Spend as much as they can on tokens and then take whatever's left except they lose too much money for that strategy to actually work.
Speaker A
And that puts every corporation in America in a very awkward situation because let's say you're the CEO of a corporation, right? You just spent $50 million on AI this year. So your board of directors asks you a question.
Speaker A
They're like, "What did we get for this $50 million we just spent? What's the ROI?" And you're like, "I don't know, right? We don't have a number. Nobody has a number. Corporate America's paying subscription fees on a technology whose
Speaker A
outcome we cannot measure. And it gets worse though because not only can we not measure it, we are also risking our company secrets and potentially creating a competitor. Here's what Alex has to say about that.
Speaker A
But something has gone completely wrong. And the basic view among enterprises in this country is I'm going to chill, lax, uh, and waste my time with tokens. I'm going to get no value and they're going to get my IP.
Speaker A
The fear for all these CEOs is that when your company uses these models, your data flows through them, right? Your process, your trade secrets, the special sauce that makes you profitable, which Alex Karp calls the alpha. So what
Speaker A
happens when the AI company that you use learns from your business? It just becomes your competitor. And this is not a hypothetical thing, by the way.
Speaker A
Anthropic launched a design product called Claude Design while having a relationship with a company called Figma, which is a design company. Figma's CEO publicly said he was shocked. So, picture being a business and then watching that. You're paying your vendor
Speaker A
millions of dollars a year to use their AI, but what you're actually doing is paying them and training your own replacement. So, what's the solution then? This is where it gets really interesting.
Speaker A
But what is happening among the most technical players is they're sayi
Speaker A
I want to control the alpha. Why would they get access to my data? If they're going to build my alpha, why wouldn't I control the weight?
Speaker A
Right? What that means is instead of renting an AI from someone, instead you just download one, you run it on your own computers using your own data where no one can see it and no one can learn from it and also no one can shut it off.
Speaker A
He then goes on to say though that most businesses don't even need the latest and greatest cuttingedge AI because you don't need the smartest one to process insurance claims, right? You need one that's specifically really good at insurance claims. And what's interesting
Speaker A
is that Alex Karp profits from this business model as well. So why would he be saying all this? What's his motive?
Speaker A
Cuz remember France, Germany, Spain, they're canceling their contracts. Palanteer has been losing those contracts across Europe all year. Now, 2 days before that interview, Palanteer announced a partnership with Nvidia to sell open models in sovereign environments.
Speaker A
Basically, that interview was a product launch for his new service, which is why he's out there telling other companies to download and own their own AI. So, that is the first problem with AI.
Speaker A
Nobody trusts it. But there's a second problem that's even bigger. Now, before I explain the second problem, everything in this video, like the AI spending and whether this is a bubble at all, depends completely on where you're reading it.
Speaker A
And that's where today's sponsor, Ground News, comes in. Ground News is an app that shows you the same story from hundreds of different outlets at the same time. And it tells you which ones are left-leaning, right leaning, or
Speaker A
center, so you can see how the story changes depending on the source. Perfect example, South Korea just announced a huge national AI and chip investment.
Speaker A
Over 159 news sources covered it. And here's what ground news shows. Left-leaning outlets frame this as a historic industrial strategy, stressing the huge scale, and they lead with the market's outcome. Right leaning outlets lean into the urgency and survival using
Speaker A
phrases like race against time, framing it as existential for South Korea's chip industry. Now, center outlets just skip the drama and focus on policy execution.
Speaker A
Even the headline number changes depending on the source. Some report a thousand trillion one, others up to 2,000 trillion, some just say 1.2 trillion, but it's the same announcement with three different narratives. That's what ground news makes visible. I use it
Speaker A
when I'm researching for these videos so I can separate what's actually happening from how it's being told. If you want to see the full picture of global events and not just one side, go to ground.news/jick or click the link down below. You'll get
Speaker A
40% off the Vantage plan, which is the one I use, and you'll be supporting independent journalism and this channel.
Speaker A
Thank you to Ground News for sponsoring this segment. And now, let's get back to it. Now, the second problem with AI is that even if every company in America trusted these AI companies, the money still does not make sense because the
Speaker A
business model is broken in a way we've never seen from tech companies. Because here's how software is supposed to make money. Software is the greatest business model ever invented because you spend a lot of money building the thing at once,
Speaker A
right? And then every new customer is basically free money. That's why tech stocks have done so well over the past few decades. When you buy Microsoft Excel, right, Microsoft doesn't spend anything extra to sell you that copy.
Speaker A
Their costs stay the same, but their revenue goes up. And it's the gap between those two lines that is their profit. And that gap is why tech companies, the most valuable companies on Earth. Now, AI broke that model. And
Speaker A
how they broke it was every single time you ask Chad GPT a question right now, it costs Soap AI money cuz it uses electricity. Chips are being worn down.
Speaker A
So more customers does not translate to free money anymore. More customers means more cost dollar for dollar. So AI is not a software business. It's more like a restaurant, right? Every time a meal gets served, somebody has to buy the
Speaker A
ingredients every time. Except this is a restaurant that loses money every time it serves food. And its plan to fix it is to serve more food. Now, let me give you some context. In 2025, Open AAI burned over $20 billion in just one
Speaker A
year. Well, they'd be the first to be this bad other than we work. And even then, this is so much worse than that. Open AAI burned $20.9 billion in 2025. That's the auditive financials that the FT and II
Speaker A
reported. And the problem with these companies is their margins are getting worse and they actually their costs increase linearly with their revenues.
Speaker A
So he's basically saying that costs increase linearly with revenues, right? The two lines are going up together and the gap never really opens up. Now for 25 years, every investor has been trained to be patient with this cuz they
Speaker A
say, "Well, they're losing money now, right? But at scale their margins will get better, right? Amazon lost money for years. Except with AI, we just keep on waiting and the margins are getting worse cuz every new model costs more to
Speaker A
run than the last one. And the market is starting to notice it. There is no proof that they can improve their margins. No amount of specialist silicon or supposed Vera reubans will bring these costs down. And we're at a
Speaker A
point now where OpenAI is now potentially pushing their IPO to 2027 because they couldn't get a trillion dollar valuation. It's clear that people are wising up to the problem of generative AI, which is there's not really a business there.
Speaker A
Now, all of this, by the way, is not just open AI cuz look at who's paying to build all of this. This is from Oracle's annual report.
Speaker A
Oracle is a particularly scary one because they are building 7.1 gawatt of capacity just for one customer. And they even said in their annual report that the risk was they might not get paid.
Speaker A
Open AI only loses money and I think I estimate it's like $75 billion of revenue annually that they will have to pay for the full Stargate data center project in annual compute revenue. Open AAI can't afford that and if they can't
Speaker A
Larry Ellison can't afford to pay back those bills and Oracle stock will be in jeopardy along with the margin loans that Mr. Ellison holds. It's genuinely dangerous.
Speaker A
So Oracle is building the equivalent of several nuclear power plants worth of electricity for basically one customer.
Speaker A
a customer that just lost $20 billion. Here's my favorite one, though. Nvidia sells its chips to a group of smaller cloud companies. They're called NeoClouds. Now, those companies borrow billions of dollars to buy Nvidia's chips and Nvidia rents them back.
Speaker A
I think companies like Core and especially Nebius and Iron and Cipher Mining and all of them, Terowolf as well, they are all very they're basically outgrowths and they're subsidiaries of Nvidia. Nvidia is now according to the information going to be
Speaker A
paying them to rent back their GPUs when they install them in the data center.
Speaker A
This is the this is something that only happens in an industry without diverse and real demand.
Speaker A
So what he's saying there is that Nvidia's sales are partially funded by Nvidia. That's like a car dealership lending you money to buy a car and then paying you to borrow the car back for the weekend and then reporting all of it
Speaker A
as demand. Now look how much profit we're making, right? Yeah, because you are buying back your own equipment.
Speaker A
There's a name for when an industry starts doing this. It's called not enough real customers. So for companies investing trillions in AI like Microsoft, Google, Amazon, Meta, what is the ROI from all their spending? They won't tell you. These companies report
Speaker A
everything. Cloud revenue, ad revenue, YouTube revenue. But AI revenue, they're not telling us that. Microsoft, Google, and Meta, and Amazon are all doing a funny little I don't want to call it a scam, but it's a a trick where because
Speaker A
their other businesses are still growing, but they never disclose their AI revenues, everyone conflates that with AI driving their growth. In reality, their other businesses are growing and AI is losing them money across the board. You'll notice that
Speaker A
neither Microsoft or Amazon, who both share their run rate of AI, will share the actual AI revenues. That tells you that these companies are afraid. Public companies love good news. If they had good news, why wouldn't they share it?
Speaker A
That's because they've only got bad news here. Now, as of right now, the stock market is still patient and investors are saying, "Okay, give it time still." But all of it really depends on one big assumption, which is that if and when
Speaker A
the profits do come, it's going to be the American companies that will make the profits because the world has no other option. But the third problem with AI is that the world has another option.
Speaker A
That option is called China. So, let me show you what China is really doing.
Speaker A
Remember this chart from the beginning of the video. The trillion that America is spending. Well, here's something interesting. This is China. On that same scale, America, $764 billion this year, and then 1 trillion next year, 3% of the
Speaker A
whole US economy. China, 102 billion this year, 123 billion next year. 0.6% 6% of their economy, which means America's outspending China almost 10 to1. Why? It's cuz China figured something out. A developer took the exact same coding task and gave it to
Speaker A
two AI models, Claude Opus, which is one of the top American models made by Anthropic, and GLM, which is a Chinese open model. Both models finished the same task, and both took about 5 1/2 minutes. The American model charged
Speaker A
$2.33 and the Chinese model charged 31. That's 7 12 times cheaper. Now, before you think that that's a cherrypicked test, here is the industrywide data. This is called the artificial analysis intelligence index. And this is basically the official rankings of every
Speaker A
AI model in the world. Now look at the top. The best American model scores 60.
Speaker A
Now look right here. This is GLM. The best Chinese open model 51. And look at how much of this chart is from China.
Speaker A
Deepseek, Quen, Kimmy, Miniax. They're not at the top. They are everywhere. They are filling the whole middle of the global rankings. Now to be fair, the US still has the smartest AI in the world.
Speaker A
And that's true. But if you ask the question that every business is asking, do I need the fastest AI model to manage my business? The answer is no. For most businesses, that's things like answering their customer service emails to
Speaker A
basically do the boring work that is 90% of what companies actually use AI for.
Speaker A
Based on that logic, China's winning. The US is winning the race for the most dollars spent, and China is sort of winning the race for the customer. Now the question is how is China doing this while spending 10 times less money? And
Speaker A
the answer is distillation. Okay, here's how it works. When you train a frontier AI model from scratch, that means spending billions of dollars teaching it everything the hard way. But there's a shortcut. You can train your model by
Speaker A
studying the answers of a model that already exists. This is basically like copying someone else's homework and then you know the answer for almost no money spent, by the way. So the US spends trillions of dollars doing the hardest
Speaker A
research in human history and then China just sort of copies it by compressing the results into smaller cheaper models and then it gives them away for free. It open sources them which means anyone can download them. And if you think about it
Speaker A
every dollar of US AI spending it's kind of like a donation to the Chinese AI industry. And this has become common practice for China. So much so that they are doing it as a side hustle. Check this out. This is a model, for example,
Speaker A
called LongCat. Look at the benchmarks. It's going toe-to-toe with Google's Gemini, and it's beating older versions of Anthropic's flagship models on realworld agentic tasks. But the thing is, do you know who built Longat? It's a company called Mtoan. Do you know what
Speaker A
Mtoan does? It's a food delivery company, right? It's the Chinese Door Dash equivalent. And they built an AI that competes with the smartest labs in the United States. Which means when a company like that can do what a trillion
Speaker A
dollar US company is doing, is that company still actually worth trillions of dollars? Maybe not. Cuz remember the assumption that's holding up the whole AI stock market is that let's be patient guys. the profits will come and when
Speaker A
they do all the US companies will collect those profits because the world has no other choice. But here's what the actual cost is when there is a choice.
Speaker A
Right? This is the same type of work. The American model shows 18 1.5 cents per task and the Chinese model 4 cents.
Speaker A
Right? Within a few quality points of each other at a 76% discount. This is sort of the chart that destroys the whole story because you cannot make back a trillion dollars selling something that your competitor is giving away at 90% of the quality for
Speaker A
just 10% of the price. So, let me tie all of this together. If all of this is true, then when does this bubble pop, if ever? And logic says it's when corporations stop their capex, their capital expenditures. It's when they
Speaker A
stop spending money building all these data centers. But believe it or not, that is not when the bubble pops.
Speaker A
According to the data, data shows that the bubble could pop a lot sooner. And here's why. During the dot bubble, the NASDAQ index peaked in March of 2000.
Speaker A
Now, all the companies that were laying the fiber optic cables at the time, which are the data centers of that era, they kept spending billions of dollars well into 2001, even though the stock market collapsed a full year before
Speaker A
their spending stopped. So, the market did not wait for companies to stop spending and to admit to anything. The logic of the market changed when enough investors stopped believing in that story. So the trigger this time around I
Speaker A
think will be something a lot more subtle. Something like a big tech earnings call where a CEO says something like we are moderating uh the pace of our infrastructure investment or some boring small thing like that. And that's
Speaker A
because the first company that gets rewarded by Wall Street for cutting their AI spending that will give every other CEO permission to do the same thing. In fact, according to Ed Zitron, Goldman Sachs recently said that the first hyperscaler to pull back on
Speaker A
spending will get rewarded by the markets. So, I heard Goldman analysts say recently that the first hyperscaler to pull capex will get rewarded by the markets. I think the capex pullbacks are they're the sign. I also think any
Speaker A
financing falling through Baro and AI or anthropic would be a sign, but I think we're going to start seeing AI companies kind of start falling out of favor and not being able to raise money. But the big thing is debt. When data center debt
Speaker A
stops being issued, that will be when it's bedtime for this industry because even if they think AI is going to win, we've got 100 gawatt or so of data center capacity allegedly under construction or in planning. That's trillions of dollars of money that needs
Speaker A
to come from somewhere and we are tapping out the debt markets. We saw that with Google raising that $85 billion equity raise.
Speaker A
That's going to be one of the early signs. Now, another sign that we could be at the peak of the bubble is the bond market. That's because unlike the stock market, which runs on stories of hopes and dreams, the bond market doesn't work
Speaker A
like that. All bond investors care about is will I get paid my interest payment.
Speaker A
Right? The moment they get scared, they start to demand a much higher interest rate. Now, how we measure their fear is something called a credit spread. Here's how that works. In the world of investing, there's a concept called the
Speaker A
riskfree interest rate. It's called that because it is set by the US government which is considered to be the safest borrower on earth. That's government bonds, right? Whatever they're at, that's the risk-free rate. Okay? But remember, companies can also issue
Speaker A
bonds. Except because companies are risky, cuz they can go out of business, their bonds pay that risk-free rate plus something extra to compensate you for the risk that their company could go broke and never pay you back. Makes
Speaker A
sense, right? Well, that extra between the risk-free rate and their rate, that is called the spread. Think of it as an insurance premium. When lenders feel safe, the premium is small, meaning the spreads are what's called tight, meaning
Speaker A
corporate bond rates are close to the risk-free rate. But when investors feel like there's some market risk, the premium explodes, right? The spread increases. That is one of the early signs that we could start to see that this is going to fall apart. Now, here's
Speaker A
an example. By the way, see this increase in 2008. Spreads hit almost 22%. Lenders started to charge very high prices. Credit shut off completely and companies that ran on borrowed money just collapsed. Also see the jump in 2020. Now look at today. The spreads are
Speaker A
very tight. 2.6%. That is close to the lowest and the calmst readings in recorded history.
Speaker A
What this means for now is that either bond investors see no problem and everything in this video is completely wrong or bond investors are wrong and they can be wrong. Look at early 2007.
Speaker A
The housing crisis was already underway. Bear Sterns was months away from blowing up and spreads were only 2 1/2%. They were super calm right where they are today. Right? The fear gauge didn't predict 2008, though. That's because spreads don't really measure what is
Speaker A
true. They measure what lenders believe. In 2007, lenders believe the housing market was safe. Which now we know obviously that it wasn't. But the point is is that when you see someone on the news say, "Hey, don't worry. AI is
Speaker A
totally safe. It's doing great. Credit markets, right? The spreads are not so worried." Right? If you hear that, remember that the credit market wasn't worried in ' 07 either. Credit markets can be wrong and they have been wrong
Speaker A
before. Although you you'd agree that spreads do not do not imply that that moment is anytime soon.
Speaker A
Spreads have been wrong before. That's the thing. And I think that perhaps the timing isn't going to be immediate, but at some point a hyperscaler is going to pull back capex. And when that happens, well, this is an industry of followers.
Speaker A
The tech industry doesn't have ideas. They just copy each other. Everyone copied Sachin Nadella when he put chat GPT in Bing. And I think that whoever breaks capex first, they'll follow them, too.
Speaker A
And finally, I just want to show you what Michael Bur posted a few days ago.
Speaker A
And remember, he's the guy who predicted the 2008 financial crisis. So, in chart one, he shows chip stocks are trading at the top of their 15-year valuation range. Basically, the same peak that they hit right before the 2024
Speaker A
correction, which is marked with those red circles. The market is basically pricing chips like the trillion dollars has already been made. Now chart two is even more interesting. This tracks the three groups of AI stocks since last year. Now the gray and white lines going
Speaker A
up to 200% are the AI winners, right? The companies selling the chips and the equipment. But the orange line way at the bottom that's barely above zero are the hyperscalers, right? Companies like Microsoft, Google, Amazon, Meta. What does that mean? It means the market is
Speaker A
telling us that the companies that are doing the spending, the trillions of dollars, right, they're getting almost no credit for it. Their stock values aren't really going up. Wall Street instead is rewarding the companies that are getting that money and it's ignoring
Speaker A
the companies spending to build it. Right? That's basically the market admitting it doesn't believe the spenders will make it back. And in the third chart he posted, it shows the Silicon Data LLM token expenditure index. It's a fancy name, but what it
Speaker A
shows is it shows us the price that people pay for AI tokens. This index is the price of AI itself. And look at it.
Speaker A
It's down almost 20% from its high in May. Now, the question is, why would the price of AI be going down during the biggest AI buildout in history?
Speaker A
Bloomberg says either it's because demand is going to cheaper models or buyers are just not willing to pay more.
Speaker A
Look at the middle one. Demand is shifting towards cheaper models. And that is the China theory that's showing up in this data. Now, to be fair though, Michael Bur's been early before. And when people say early in the market,
Speaker A
that's a polite way of saying he's been wrong, right? He's made market crash predictions over the years quite a lot.
Speaker A
That didn't really come true. And even this index has dips that have recovered. Basically, Bloomberg says that the signal for all of this is ambiguous, right? We can't really learn anything from this data. It can mean anything.
Speaker A
So, basically, the real answer to how long it will take for the AI bubble to pop, if ever, is that no one knows. But those are some of the early signs to look for based on the data from the
Speaker A
past. Now, if you're interested in seeing how I'm personally preparing and more of my thoughts, those videos live in the premium member section where you'll also get access to my videos earlier. And if that's valuable, the link is down below. It allows me to make
Speaker A
more videos like this one and take on fewer sponsors. Thank you for watching and being a premium member. I hope you have a wonderful rest of your day. Smash the like button, subscribe if you haven't already. Would love to see you
Speaker A
back here next time. See you soon. Bye-bye.
Topics:AI bubbleChina AIUS stock marketAnthropicPalantirAI investmenttechnology competitioncorporate AI risksAI ROItech bubble
