Why investors buy into AI while calling it a bubble

HostIt's such a strange time to look at the news. One day you see a huge report from a big bank saying AI is a giant money pit that might never pay off, and the next day, those same firms are helping tech giants spend billions more on it. It feels like everyone is shouting that the house is on fire while they keep running inside to put more furniture in the living room.

HostWhat's actually going on in the heads of these investors who seem to be saying one thing but doing another?

GuestIt really does look like a total split in the brain from the outside. But the thing to understand is that for these big tech firms, the danger of being too slow is way worse than the danger of spending too much. They're basically in a race where the track is still being built as they run on it. They're spending all this cash on the plumbing right now. I mean the high-end chips and the massive data centers that take years to build. If you don't have that stuff ready, you can't even play the game when the next big breakthrough happens. So even if they're worried the hype is too high, they feel they can't afford to sit out and let a rival win the whole thing. It's a bit like a game of musical chairs where the music is still playing, but every chair costs ten billion dollars.

HostBut if we don't know if this is going to make money yet, isn't that just a classic bubble? We have seen this before with the dot-com era where people threw money at anything with a website.

GuestThere are definitely shades of that. The big worry right now is what people call the ROI gap. That's just a fancy way of saying the bill is coming due, but the paycheck isn't in the mail yet. We can see the cost very clearly. We see the power bills and the chip orders. But we're still waiting for the killer app that makes regular people pay for this every month. Right now, a lot of the money being made is just tech companies selling stuff to other tech companies. It's a bit of a closed loop. If that loop doesn't break open and start solving real problems for normal businesses or people soon, then yeah, the bubble talk gets a lot louder.

HostSo why not just wait? If I'm a smart investor, why wouldn't I wait for the dust to settle and see who actually wins before I dump my life savings into it?

GuestWell, that's the catch. In the world of software, being first often means you win everything. If one company builds a tool that's twice as good as everyone else's because they had more computing power, they might capture the whole market before you even get your boots on. Investors are terrified of missing the next Google or the next Apple. They would rather lose some money on a bubble than miss out on the thing that changes the world for the next thirty years. It's less like a typical stock market gamble and more like a land grab. You grab the land now because they're not making any more of it, even if you're not sure what you're going to grow there yet.

HostThat sounds like a lot of pressure to put on a guess. I mean, we're talking about hundreds of billions of dollars. If this really is just a bubble and it pops, doesn't that take down a huge part of the economy with it?

GuestIt could be messy, but here is where it gets interesting. Even if the stocks crash, the stuff they built stays there. Think back to the late nineties when everyone was laying miles and miles of fiber optic cable under the ocean. People called it a total waste. The companies doing it went bust, and investors lost their shirts. But those cables stayed in the ground. They were what made the high-speed internet possible a few years later. They were the reason we could've things like video streaming and social media. So, even if we're in a bubble, the huge data centers and the new power grids being built right now are real things. They'll be there regardless of what the stock price does tomorrow.

HostSo the waste might actually be useful later on? That's a bit of a silver lining, I guess. But it still feels like we're ignoring the fact that these tools are incredibly expensive to run. A search that uses AI costs way more than a regular old search.

GuestIt does. And that's a huge point of friction. Right now, every time you ask one of these bots a question, it might be costing the company a few cents in power and hardware wear and tear. That adds up fast when millions of people do it. The big bet is that we'll find a way to make it cheaper. History says we usually do. Computers used to fill whole rooms and cost a fortune, and now we have them in our pockets. But there's no rule that says it has to happen this time. We might hit a wall where it just stays expensive, and that's where the real crash would happen.

HostIt seems like we're all just waiting for that one moment where the tool becomes more than just a fun toy and starts doing the heavy lifting for us.

GuestThat's the big question. We're waiting for the moment when a company can show they saved a billion dollars by using this, or a small business can do ten times more work with the same number of people. Until that happens, the people writing the checks are mostly living on hope and the fear of being left behind.

GuestThe real test is whether we're building a foundation for a new city or just building the world's most expensive sandcastle that the tide is about to wash away.

HostThe high-speed cables under the sea are still carrying our data today, even if the people who paid to put them there never saw a dime back.