원문 보기
It's the largest AI infrastructure project by far in history. I think AGI is coming very, very soon. How can a company with 13 billion in revenues make 1.4 trillion of spend commitment? want to sell your shares, I'll find you a buyer. AI and semiconductor stocks tumbled, shedding more than 1 trillion dollars in value this week.
In the first signs of what feels to me like unsustainable silliness. Why should we trust you? You shouldn't. Sam Altman's company is worth 852 billion dollars. It has lined up 1.4 trillion dollars of spending on chips and data centers. Last year, it brought in 13 billion dollars. But there was one problem.
It lost 21 billion dollars running the business. We have no current plans to make revenue. We have no idea how we may one day generate revenue. And this week, the money that pays for it all got expensive. This is the story of Sam Altman's 1.4 trillion dollars bet, and why the man who started the boom is now warning about it.
In 2005, a 19-year-old dropped out of Stanford to build a phone app. I started a company that went through the very first batch of Y Combinator. Three years later, he was on Steve Jobs' stage. We are incredibly psyched about LoopChat on the iPhone. The app struggled. In 2012, it sold for 43 million dollars, about what it had raised.
But at 28, he was running Y Combinator, the startup school behind Airbnb. Generally, my view on like all of life is that things are not as risky as people think they are. In 2015, he joked on stage about where AI was heading. You know, I think AI will probably, like most likely, sort of lead to the end of the world, but in the meantime, there'll be great companies.
That December, he and Elon Musk launched a nonprofit AI lab. Well, what would be a an open-source nonprofit because Google is closed-source for-profit. They called it OpenAI. Backers pledged 1 billion dollars. By 2019, only about 130 million dollars had actually arrived. In 2019, the lab needed far more money.
So, the non-profit created a for-profit company. Investors could make 100 times their money. He told investors how they would get paid. Basically, we will ask it to figure out a way to generate an investment return for you. It sounds like an episode of Silicon Valley. It really does. I get it. I I You can laugh.
That summer, Microsoft invested 1 billion dollars. Then, on November 3rd, 2022, the lab released ChatGPT. We shipped ChatGPT as a low-key research preview, and that went pretty well. In 5 days, it had a million users. In 2 months, about 100 million. Microsoft put in a reported 10 billion dollars more.
Overnight, Sam Altman became the face of AI. He even told Congress what could go wrong. I think if this technology goes wrong, it can go quite wrong. Then, one Friday in November, his own board fired him. Five months earlier, he had said this on stage. The board can fire me. I think that's important.
The tech world has thrown has been thrown into chaos over the weekend when the company that gave us ChatGPT fired its CEO. The board said he had not been consistently candid with them. It totally crazy and surreal and felt like a dream. I I was caught completely off guard. Microsoft, his biggest backer, got less than 15 minutes of warning.
By Monday, he and Greg Brockman were joining Microsoft instead. Then, more than 700 employees threatened to quit and follow him. And like 700 you know, like 95% or something of the OpenAI employees had signed this letter while I was asleep saying like, you know, this is a horrible decision. This was a coup of of the board over the weekend where they said, "We're going to make sure that we're doing this in the best interest for humanity." And the investors pushed back and said, "Hold up."
Five days after he was fired, Sam Altman was back in charge. OpenAI posting on X that Sam Altman will now officially return as CEO. Most of the board that fired him was gone, and now nothing could slow him down. Weeks later, The Wall Street Journal reported he was seeking trillions of dollars. The only question was who would hand it to him.
On January 21st, 2025, his first full day back, President Trump made an announcement. Share a new American company that will invest 500 billion at least in AI infrastructure. And you told me, "Oh, Masa, go off of 200." Now I came back with 500. I think this will be the most important project of this era.
They called it Stargate. The first site was already rising in Abilene, Texas. Each building is a half a million square feet. There are 10 buildings currently on currently being built, but that will expand to 20. Then the deals came, one after another. Oracle signed a $300 billion deal to sell it computing power.
Its stock jumped 36% in a day. Nvidia said it would invest up to $100 billion. This partnership, I mean, this is, you know, monumental in size. AMD gave it the right to buy up to 160 million AMD shares for a penny each. It promised Microsoft $250 billion for cloud computing. Amazon, $38 billion. In November, Altman put the total at about $1.4 trillion over 8 years.
If we knew how to get a trillion dollars right now, which we don't, would we be able to deploy that profitably in the next few years? Investors did not wait to find out. In 3 years, the company went from $29 billion to $852 billion. But critics saw a catch. Much of the money was moving in a circle. Nvidia invests in Altman's company.
The company buys Nvidia's chips. Oracle buys Nvidia chips, then rents the computing power back to it. AMD hands it shares. It buys AMD chips. People say it was circle. The merry-go-round of money continued. The largest tech companies in the world are purchasing this infrastructure because they have demand.
There's nothing circular about that. To build it, its partners were on course to borrow almost a hundred billion dollars, the FT reported. Then, at a Wall Street Journal event, the CFO said this. The the backstop, the guarantee that allows the financing to happen. There's some federal backstop for chip investment.
Exactly. To critics, a government backstop meant taxpayers would carry the risk. The next day, the White House's AI chief wrote, "There will be no federal bailout." She said the company was not asking for one. Days earlier, a podcast host had asked Altman the obvious question. How can a company with 13 billion in revenues make 1.4 trillion of spend commit
If you want to sell your shares, I'll find you a buyer. I would love to tell them they could just short the stock, and I would love to see them get burned on that. The company's own forecast showed a 74 billion dollar operating loss in 2028. Something magical has to happen in the course of 24 months.
The bet only works if the money never stops coming. Then, this week, the money got expensive. The warnings had started from the inside. And and if and if and if we end up misspending a couple of hundred billion dollars, and then it could just be impossible to fulfill the kind of the the compute funding like the compute application.
Last month, Altman himself said it. But I am seeing the first signs of what feels to me like unsustainable silliness. People claiming that they're going to build gigantic amounts of compute next year that I think they don't have the revenue to support or a buyer. Michael Burry, who you probably know from The Big Short, is now calling AI a bubble.
Matter of fact, I can you and I can find you a hundred companies that, you know, were worth 50 billion dollars and disappeared. Then came this week. CPI halt, PPI hotter, coming in more than a percentage point above expectation. The Fed had already raised rates in September, its first hike since 2023.
Hotter CPI forces the Fed to react, hiking interest rates. That means AI firms have to pay more for funding. And Coreweave, which rents AI chips to Altman's company, cracked. Its shares fell 15% in a single day. Coreweave story has gotten worse this morning with its 2032 bond remaining above 15%. That's considered by the credit market as distressed debt levels.
And over a 5-year duration, what that math implies is about a 55% chance of a default. So, it's a coin flip whether they default or not. On Bloomberg, the anchors could hardly believe the number. Over the last 6 days, the SOX has erased over $1 trillion in market cap. That's just wild. In South Korea, home of the world's memory chips, the market fell 7 days in a row.
Rising prices, a Fed raising rates, a tech boom built on promises. Wall Street has seen this before. The plunge was sparked by jumping US inflation, the fear a bigger hike in interest rates the only remedy. It lost a quarter of its value in just 1 week. John Chambers ran Cisco, the giant of that boom, through the crash.
Will there be some spectacular train wrecks and affecting the ecosystem? Absolutely going to occur. So, is this the top or just a dip? The bulls say the demand is real. The rate issue, I get the nervousness. I'm focused on demand, and demand's telling you we are third inning relative to this AI game.
But but the lack of demand killed it. We don't really have a lack of demand in AI. Uh I do think like for sure if you get into a pathological debt situation, that can wreck anybody. The bears say demand is not the same as profit. Is that there is no path to profit right now for AI. And here's why it matters, even if you never bought an AI stock.
Whether they realize it or not, it is rolled into our 401Ks, mutual funds, and other indirect purchases. Five tech giants now make up nearly 30% of the S&P 500. Sam Altman says he is not worried about his own plans. Am I worried about our compute build-out plans? I am worried about the world's compute build-out plans.
You know, you can imagine a world where there are some people that made dumb financial decisions. That happens in kind of like every boom or most of them. 12 years ago, he said life is not as risky as people think. Now, $1.4 trillion depends on him being right.
Comments 0