Ed predicts OpenAI could run out of cash in 2027

Transcript

From the transcript The Diary Of A CEO publishes with the episode.

1. 2:08:27

   So the big thing that people say is, oh, we'll get bailed out. Donald Trump, scared of Donald Trump. Here's the problem with this. It isn't just an AI bubble, it's the Rotcom bubble. So the AI bubble collapsing will probably be this company running out of money. OpenAI. And the thing is with OpenAI is they were meant to go public this year. And now it's been pushed to next year. A week and a half after I released their auditive financials.

2. 2:08:52

   I wonder where that was. But they've delayed to next year. Sarah Fryer, the CFO, has now said... Well, they'll do it earlier than 2027 or 2027. Great answer there. For anyone that doesn't understand what going public means, that means joining the stock market. And at such a time when you join the stock market, your investors can finally sell their equity that they got for investing in the company when it was private.

3. 2:09:14

   So oftentimes companies will flirt with the idea of we'll go public someday soon because investors will have a moment in their head where they'll get their money back at a return. So you kind of need to, if you're in these guys' shoes, you kind of need to be flirting with going public or investors won't want to invest. OpenAI up until this point has been a private company. And their last funding round, they were valued at $865 billion.

4. 2:09:38

   Now, when they tried to go public, New York Times' Mike Isaac reported this. They tried to list, well, they wanted to go at a $1 trillion valuation. Apparently, their advisor said, no, don't do that. That is very bad for a number of reasons. One, OpenAI needs perpetual amounts of money. They raised $122 billion this year. Most of it's crossed. There's some left. But... They are going to need to raise at least $100 billion a year just to survive.

5. 2:10:05

   If they can't go public, they will have to raise another funding round. The problem is it's going to be difficult to raise even the same one they raised that. They're probably going to have to take a flat, so the same amount. They can't do that. Exactly. But they need money. They need money so bad. Amazon sent them $35 billion that was meant to be contingent on them going public early. They did that because they need the money.

6. 2:10:27

   Now, OpenAI is the kind of catastrophe center here because Anthropic is likely going to beat it to go public. And once Anthropic goes public, it'll be borderline impossible for OpenAI to do so because Anthropic, an unprofitable, unsustainable AI lab, but a better business that's growing faster than OpenAI's. I believe they have a ceiling. They're eventually going to face perdition too. I think sometime in 2027, things are going to start running out of steam.

7. 2:10:49

   Because the thing I said earlier, the only way these models get better is if you feed more money, tens of billions of dollars into them. So you think OpenAI runs out of steam in 2027? I think they're already running out of steam, yeah. But I think they run out of cash. You think they run out of cash? Yes. And the sequence of events here will be they go out and try and raise... And they have trouble raising another round.

8. 2:11:08

   I think maybe NVIDIA props them up a little. Maybe Private Credit, Blackstone, BlackRock, and the like. And the reason that Private Credit is getting involved, so asset managers, is because they're investing in the data centers, and they know this company's most of the data center demand. Okay, so they run out of steam in 2027, according to you. Yep. And maybe they try, if they bum rush to go public, they're going to have worse economics than an anthropic.

9. 2:11:28

   They're going to get savage. WeWork was a great example. Another SoftBank classic. Now, I think OpenAI collapses. There are many different ways it could happen. There are many different ways it could end. But the crucial thing is that there are multiple companies that are existentially tied to OpenAI. SoftBank, one of the largest companies in the Japanese stock market, a holding company with lots of investments, they have on paper...

10. 2:11:52

    about $100 billion worth of OpenAI stock. If they can't go public, they can't do diddly squat with them. And so SoftBank's future, their ability to continue paying the people around them and existing as a business, relies on their ability to continually liquidate funds, to take the things they've invested in and have value from them, either by selling the stock or taking loans out on the stock. If OpenAI can't go public, SoftBank can't do that.

11. 2:12:17

    SoftBank. Probably won't run out of money, but we're going to see one of the largest holding companies in the world become much smaller. We will also see Amazon, Google and Microsoft have to restate guidance. They will have to say, actually, we don't think we're going to grow as fast. And what happens then? Well, I think we enter a tech depression because the rock-com bubble, the core of my theory, is that they're out of hyper-growth ideas, but the market doesn't think so.

12. 2:12:41

    The reason they're so maniacally spending is because buying AI GPUs allows them to kick the can further. It allows them to say, we're still doing something. We're working on AI. Don't think too hard. And also, their current business is still growing. Yeah. their current businesses will eventually slow. There's only so many price increases. There's only so many tweaks to ads. Only so many tweaks to Google search.

13. 2:13:03

    Only so many ways that Amazon can screw merchants. So in that tech depression, which you think it might be triggered in 2027, is that a cascading downstream economic depression? Because the stock market is heavily dependent on these companies. The stock market sees a pullback. Investors stop investing. They get panicked. Yes. I think that because... What sort of downstream consequence, the sort of domino effect?

14. 2:13:27

    There's so much to imagine that it's difficult to capture everything, but there are a few things that worry me. First of all, a ton of American money, just regular people's money, retail investors, are in these companies. And they bought into the Magnificent Seven thinking that number go up forever. NVIDIA is the largest company on the Fortune 500 and NASDAQ as well. And like 7% to 8% of the S&P 500...

15. 2:13:49

    That company, when the bottom falls out from NVIDIA, and we haven't really got into it, but NVIDIA is doing the most circular of financing, feeding companies money so that they can raise debt to buy more GPUs. I think NVIDIA's revenue could go 50% to 70% now. I think that NVIDIA could, NVIDIA back in 2022 was making single digit billion dollars. And what happens though, I'm thinking about like Jenny and Dave that are watching this right now, and they are just normal people.

16. 2:14:12

    With normal jobs. People's retirements are going to contract severely, and I don't believe they're going to return to those values. And I think that because so much of the value of the S&P 500 and Russell 1000 index comes from these four companies and the rest of the Magnificent Seven, so Apple, Tesla, Meta as well. And the thing is, I don't know what happens after that because venture capital has also, like more than half of venture capital last year went into AI.

17. 2:14:38

    I think most venture capital investments in AI are going to zero. Because when it comes to building a company on top of an LLM, all of those are unprofitable too. And the thing is, LLM companies have not really been acquired, the exception being Cursable by Elon Musk for the coding side. But you have Cognition, which is just another LLM company raising a $26 billion valuation. That means that company has to go public because who's buying a company at $26 billion other than Elon Musk?

18. 2:15:04

    And there are rumors that Elon Musk was trying to buy them as well. Is Elon Musk just going to pick off every LLM? Like LLM company, like on a fucking TJ Maxx for AI, like Jesus Christ. So is that a recession you're describing? It is a recession, but it's also a depression within people's retirements. Like I'm talking about 20, 30, 40% off the top of these companies' stock value. Economic contractions, recessions consistently lead to job losses and rising unemployment.

19. 2:15:27

    When an economy contracts, the mechanism driving job losses typically follows a predictable sequence.