I just saw Nasdaq drop sharply again, and some people are starting to shout "AI bubble." This made me think of a very interesting historical comparison—the dotcom companies during the internet bubble era.

Honestly, the stories of those internet companies back then are very similar to today's AI narrative. In the mid-1990s, the internet transformed from a niche technology into something every household could use, and capital went crazy. Venture capitalists rushed to throw money, as long as you dared to call yourself a dotcom company, investors would pour money in. A very typical phenomenon at that time was: as long as a company added ".com" to its name, its stock price could double. No need to care about profits or a real business model—just "believe in growth."

By around 1998, this madness reached its peak. Nasdaq was almost rising vertically, with waves of internet companies going public, doubling on their first day. Media outlets praised young entrepreneurs daily, stories of dorm-room startups turning into millionaires were everywhere. Retail investors started to frenzy over stocks, abandoning basic risk management, all-in on these dotcom companies. The entire market logic was: as long as the internet kept expanding, traditional financial metrics didn't matter anymore.

But the problem was, most of these internet companies were burning money like crazy. They needed continuous funding to expand, do marketing, build infrastructure, but they couldn’t see the day of profitability. Many analysts pointed out this contradiction at the time, but no one listened. Everyone thought "this time is different," that the internet would rewrite economic rules.

What happened in reality? In early 2000, the Federal Reserve started raising interest rates, liquidity tightened. Some large tech companies announced disappointing earnings. That burst the "internet is omnipotent" bubble. From the peak in March 2000, Nasdaq plummeted nearly 78% over the next two years. Those dotcom companies once seen as "representatives of the internet revolution" either went bankrupt or their stock prices became worthless.

Cisco was a good example. It once became the world's most valuable company, with its stock soaring to $82. But then it kept falling, and to this day, it hasn't regained that high. Thousands of startups disappeared, Silicon Valley office buildings went empty, countless people lost their jobs. This crash wiped out trillions of dollars in market value.

But here’s an interesting twist: although most dotcom companies died, those that survived—like Amazon and eBay—later changed the world. Their survival was simple: they had real cash flow, genuine business models, and focused on profitability rather than just growth.

Looking at AI now, I think some similarities are worth warning about. The market enthusiasm for AI is indeed very high, and companies like Nvidia do have strong cash flow and real demand support. But I’ve heard "this time is different" too many times. The internet was truly revolutionary back then, but that didn’t mean every dotcom was worth investing in, nor that valuations could inflate infinitely.

The key difference lies in fundamentals. Nvidia is indeed making money now, has pricing power, and real demand for its products. But if market expectations start to diverge from actual profitability, that’s dangerous. History shows us that no matter how great the technology, it can’t change this fact: sustainable cash flow, operational efficiency, and real profits are what determine long-term value.

Ultimately, investors’ psychology hasn’t changed in decades. FOMO, herd mentality, being fooled by stories—these repeatedly push asset prices to irrational heights. The dotcom bubble is the best lesson. Bubbles may burst, but truly innovative companies will survive. The question is, before the bubble bursts, how do you tell which are genuine innovations and which are just hype? There’s probably no simple answer.