Is AI a bigger bubble than the dot-com crash?

Transcript

In March 2000, the most valuable company on Earth was Cisco, which sold the hardware the internet ran on. It made $2.7 billion in profit that entire year. In Vidya, sales the hardware AI runs on, and it now makes more than 10 times that in a single quarter. So when people say AI is a bigger bubble than .com, they are comparing the wrong thing. The question was never the price. It is who is paying. In 2000, investors and lenders were paying for companies that mostly made no profit at all. When the money stopped, the companies vanished. Pets.com lasted nine months as a public company. Cisco itself fell 86% and has never regained its 2000 peak. This year, Amazon, Microsoft, Google and Meta are spending roughly $700 billion mostly on AI. And yes, it is eating almost all their spare cash. Google's free cash flow turned negative last quarter. But if AI disappoints, they do not vanish. They stop building. And they go back to selling ads, cloud and software. Two billions of people, the way they did before any of this started. That is the difference. In 2000, the bubble was built on borrowed belief. In 2026, it is built on the profits of the four richest companies in history who can afford to be wrong. So AI stocks can absolutely crash. The company's paying for AI will not die with them.

Transcribed automatically from the video, so small errors are possible.