Opinion: AI will change the world but that doesn't make it a good trade

Every time someone says “the AI trade is getting crowded,” some people immediately act like you are denying AI exists or that you missed the rally and now you are bitter.

That is not the point at all.

AI can be real, transformative, and still be a bad trade at the wrong price. People confuse those two things constantly. Being right about the technology and being right about the stock are not the same thing.

Let me explain with a simple example:

During the dot com era, a relative of mine bought Nortel close to the top. And honestly, his basic thesis was not even wrong. The internet was real. It did change the world. Everyone knew it was going to matter. The problem was that by the time everyone agreed, the stock was already priced like nothing could ever go wrong.

So he was right about the future and still lost most of his money. Similar example with AT&T as it took that stock 20+ years to recover from the bust.

That is the part people do not want to hear. You can understand the future correctly and still buy the wrong asset, at the wrong price, at the wrong point in the cycle. I think AI trades are too much too soon.

Not because AI is fake. Not because Microsoft, Google, Amazon, Meta, Nvidia, Broadcom, etc. are fake companies. The demand is obviously real. These companies are spending real money. Data centers are being built. Chips are being ordered. This is not some random penny stock pump.

But that is exactly what makes it dangerous. The story is now so accepted that nobody thinks they are taking a risk anymore. Everyone thinks they are just buying the obvious future. And usually when everyone agrees on the obvious future, a lot of the upside has already been pulled forward.

Now suddenly you have all these AI IPOs coming.

To me, that is not automatically bullish. That is usually when insiders start selling the dream to the public. Venture capitalists and private owners do not rush to IPO when they think valuations are terrible. They do it when public markets are willing to pay stupid multiples for the current theme.

That is usually how these cycles work. First the obvious winners run. Then the suppliers run. Then every adjacent company adds AI to the pitch deck. Then the IPO window opens because retail and institutions are desperate for more exposure. By that point, a lot of the easy money has already been made by somebody else.

The macro setup also matters. AI requires insane upfront spending today for profits that are supposed to show up years from now. If rates stay higher for longer, the market may eventually stop being patient with “trust me bro, the profits are coming later” stories.

The semiconductor side is probably where this gets most fragile. If Big Tech is overordering custom chips because nobody wants to be caught short, that can look like endless demand on the way up. But if end demand slows even a little, the hit to suppliers can be much bigger than people expect. That is how these supply chains work. Small change at the customer level, big pain at the supplier level. The bullwhip effect.

That is why I am skeptical of names like $AVGO at these valuations. Great company, real business, real demand. But the stock is being treated like the growth curve can only move in one direction forever.

And that is rarely how markets work.

Again, I am not saying AI is fake. I am saying the trade is crowded. There is a difference. The internet was real too. Railroads were real. Cars were real. Smartphones were real. Plenty of people still lost money buying the wrong companies at the wrong time because the narrative became too obvious.

So when I see a wave of AI IPOs, I do not see confirmation that everything is healthy.

I see people who got in early trying to sell the story to people getting in late.

Maybe this still runs for a while. Manias usually go further than reasonable people expect. But at this point, I think the risk reward is way worse than people want to admit.

The question is not “is AI real?”

The question is: if everyone already knows AI is real, who is left to buy the story at a higher price?