Transcription
Tech companies are borrowing hundreds of billions of dollars to build AI. But here's the part they're not saying. They're betting money they don't have on demand that doesn't exist.
In the same month that analysts have been sounding the alarm on the AI bubble, big tech just borrowed a record amount of debt to fund their AI infrastructure buildout. Amazon is raising $15 billion. Google raised 25 billion in debt in the US and Europe earlier this month. Meta issued 30 billion in corporate bonds in October. This came after they already raised 27 billion in debt. Oracle, who already has 104 billion in debt, is tacking on another 38 billion in private debt. This brings debt issuance to a record of more than $6 trillion this year. Tech is now financing this AI boom with borrowed money. While at the same time, we see people like Michael Bur and Peter Thiel dumping their stocks.
Nvidia recently released their earnings, and at first, everyone celebrated. According to Nvidia, they had a record revenue of 57 billion, up 62% from a year ago. After this announcement, Nvidia stock shot up more than 5%. But Michael Bur, the guy who predicted the 2008 financial crash, called out Nvidia on X. He claims the company is making their earnings look better than they actually are by dragging out the depreciation of their GPUs. This is part of a larger pattern of AI circular deals happening in big tech with Open AI at the center of all of them. Michael Bur and even Open AAI's own investors are questioning how Open AI is going to pay for these billion dollar deals when they're losing money every month.
So, I think the single biggest question I've heard all week and and hanging over the market is how, you know, how can a company with 13 billion in revenues make 1.4 trillion of spend commitments, you know, and and and you've heard the criticism, Sam. First of all, we're doing well more revenue than that. Second of all, Brad, if you want to sell your shares, I'll find you a buyer. I I just enough like But going back to Michael Bur's barrage of tweets, his last question is my favorite. One more. Open AI is the lynch pin here. Can anyone name their auditor?
Hedge funds and other institutional investors have unloaded more than $67 billion in 2025. Professionals with insider information, data, and analysis tools are pulling out. In fact, the first week of November marked the biggest net sell-off of tech shares in 2 years. But regular investors keep investing. Big tech holds all the value in the stock market. When they can't repay their debts and the AI bubble starts bursting, who do we think is going to foot the bill? Governments are already calling AI a national asset. Just like the banks in 2008, it's already too big to fail. Thanks to Open AI, we recently learned that these AI companies are talking with the government about having their debt covered if they default on the loans. But even if they're bailed out, retirement accounts will still be wiped, jobs will be lost, and houses will be foreclosed on. At the end of the day, it's always us taxpayers paying for every financial crisis.
Michael Bur claims the true in demand for AI is ridiculously small. So, I wanted to understand, is there even demand for all this AI infrastructure they're financing? If we listen to multi-t trillion dollar companies like Google, of course there is. In a recent all hands meeting, Google's head of AI infrastructure told employees that the company has to double its serving capacity every 6 months in order to meet demand for AI services. Is this actual demand or are they just scaling up to add AI into all of their products? Kind of like how YouTube offers an AI summary of a video or Gmail wants to give you an AI summary of an email. AI summaries for everything.
But according to McKenzie and company, most organizations are still just experimenting. Nearly twothirds of respondents say their organizations haven't even begun scaling AI across the enterprise. An IBM study found that over the past 3 years, CEOs say only 25% of AI initiatives have delivered expected returns. Only 16% have scaled enterprisewide.
Open AI claims there's massive demand for AI, but their users are primarily ordinary people, not businesses. In a leaked document, we can see OpenAI strategy versus Anthropics. Anthropic has focused on selling to businesses and plans to break even by 2028. Meanwhile, OpenAI is selling primarily to ordinary people, most of whom are using the free account. Open AAI has agreed to over $1.4 trillion in AI deals, and they have 800 million weekly active users. For OpenAI to just repay this $1.4 4 trillion. They would need all 800 million people to pay a onetime fee of $1,750. Obviously, that's an oversimplification, but it highlights just how much money they need for these deals that they don't have. And that's assuming that all 800 million people would be willing to pay a onetime fee and wouldn't just switch to something like Google's Gemini, which by the way, OpenAI is already losing web traffic to.
Even Enthropic's own research shows companies aren't adopting AI. They say AI adoption among US firms has more than doubled in the past 2 years. Well, it went from 3.7% in fall of 2023 to 9.7% in August 2024. They claim this shows rapid growth and that we're still in the early stages of AI adoption. But then they acknowledge that the vast majority of US firms don't report using AI in their production processes. Even more damning, the US Census Bureau tracks AI usage across 1.2 million US businesses. Large enterprise adoption of AI is actually declining. Every company size category has dropped since July except companies with one to four employees.
Is it any wonder that corporations aren't embracing AI? I mean, remember when McDonald's tried to put AI in the drive-thru? It was a disaster. You added bacon to my ice cream. I don't want I bacon. Your total is 762 at the pay window. Thank you. I'm not done. What can I get for you? Can you just start over? You'd like to start over? Yes. Okay. What can I get for you? Can you remove everything from the menu? Oh my god, what is happening? Why are there five McDoubles?
The logic here seems to be everyone needs AI because everyone is getting AI. They're building all of this AI infrastructure for demand that just doesn't exist. The only real AI revenue is happening within the tech ecosystem itself. They're just financing each other and sharing revenues.
Now, some argue technology takes time to spread. Electricity took over 30 years to reach farms after urban households. They compare it to how long it took for every household to have a personal computer. But that's comparing apples to oranges. Today, we don't need to wait around for power lines or internet cables. We're the most connected we've ever been and we already have access to AI. ChatGpt became the fastest growing app ever, reaching a 100 million users in just two months. While some of us may talk to Chad GPT like a best friend, the research shows that companies can't find the return on their investment. Many use cases feel like solutions in search of a problem. Plus, they're deliberately ignoring the social and environmental costs. Data centers are increasing everyone's electricity bills. They're consuming enormous amounts of water. They're extracting minerals from an already depleted Earth. All while companies are laying off employees claiming that they're now AI first. When technologies offer clear immediate benefits without massive downsides, they spread rapidly in our connected world. The fact that AI isn't spreading in the corporate world despite all of this AI hype and investment shows a fundamental problem with the technologies value proposition.
When I talk about tech, it's hard not to bring up politics and wealth inequality. Tech has created billionaires who are now infiltrating our government. Representatives on both sides have greatly benefited from these tech corporations. Ironically, these same representatives control merger and acquisition approvals, determining just how big tech can get. When these corporations go to build a data center in a new city, it's our representatives, the people we elected to protect our interest, who are doing private negotiations with billionaires. It's our representatives in every state approving these electricity price increases. And it'll be our representatives that approve the bailout when these AI companies can't pay for their debts because they're now considered a national asset.
Meanwhile, our government tells us that there's no money for SNAP benefits, that Social Security is running out, and that they're reducing the coverage for Medicare. All of this is our money. We pay into the system. We, the people, are funding a huge portion of this AI infrastructure, yet we own none of it. We get none of the profits. But when they fail, they'll socialize the losses, and all of a sudden, there's money to pay for it. We're already seeing AI's effect on society, and it's not great. But they've got their foot on the gas. People can't afford everyday life, rent, groceries, health care. But the rich have insulated themselves. They have no idea how most people are living. We have two economies today. We have one that is intensely market driven. For those at the top, which is extraordinary. They are spending like there's no tomorrow. Meanwhile, for those at the bottom, they're not sure how they're going to make it through the day. You talk a lot left and right as being the biggest divisions in America. It's not. Our biggest division is up and down and no more so than measured in terms of the confidence people feel. How about that? It's not left and right. It's up and down. It really is us versus billionaires. And if the billionaires are in the governments, then the governments aren't working for us.
Because at the end of the day, the people truly benefiting from AI aren't ordinary people. It's these corporations. There's times I wish I could be naive like this person believing in the narrative told to us by these tech CEOs. AI will take our jobs so that we can live in a labor-free society. A society where health care and housing isn't tied to our employment. We need only to look at history to see how this goes. Sure, they didn't have the technology that we have today or will have in the future, but people in power have been exploiting workers all the same. They're not building AI to benefit society. They're building AI to further enrich themselves with the ultimate goal that one day it will replace all of us. And they're using our money to do it.