The Nasdaq Composite closed at 5,048 on 10 March 2000. Thirty-one months later it closed at 1,114. Along the way it erased about 78% of its value, took thousands of companies down with it, and taught a generation of investors what the word bubble means in practice.
The uncomfortable part of the story is that the thesis was right. The internet did change commerce, media and communication exactly as its enthusiasts predicted. Being right about the technology and wrong about the price turned out to cost the same as being wrong about both.
How the web became investable
Until the mid-1990s the internet was an academic and military network. Netscape Navigator, released in 1994, made it usable by anyone with a modem, and Netscape's own 1995 listing — a company with minimal revenue that doubled on its first day — showed Wall Street what the public would pay for a story about the future.
The reasoning that followed was not stupid, just incomplete. The internet would be enormous. Companies built on it would therefore be enormous. Whoever got there first would own their category, so growth mattered more than profit and market share mattered more than margin. "Get big fast" was the slogan, and it was repeated by founders, bankers and analysts alike.
What went missing was the step between "this industry will be huge" and "this company will capture it at this valuation." Almost nobody was doing that arithmetic, and the people who were had stopped being invited on television.
The valuations, and the metrics invented to justify them
Between 1998 and 2000 the market stopped pricing businesses and started pricing narratives. An IPO could double or triple on the first day of trading. Companies with a few million dollars of revenue and no profit reached billion-dollar valuations. The Nasdaq went from roughly 1,000 in 1995 to over 5,000 in March 2000 — a gain of around 400% in five years.
Because traditional metrics gave answers nobody wanted, new ones appeared. Firms were valued on eyeballs — unique visitors — or on mindshare, or on revenue multiples applied to revenue that did not exist yet. Price-to-earnings ratios of 100 or 200 were common among companies that had earnings at all; a great many did not, which made the ratio undefined and, in the logic of the time, therefore not a problem.
The sell side had a conflict it did not disclose loudly. Banks earned fees underwriting these listings, and their research analysts covered the same companies. Buy recommendations on businesses with no route to profitability were not an accident of judgment.
Two companies that became shorthand
Pets.com raised $82.5 million in its February 2000 IPO. It sold pet food online at prices below what shipping cost, spent heavily on advertising including a Super Bowl commercial featuring a sock puppet, and liquidated in November 2000 — nine months after listing. The sock puppet outlived the company as a cultural reference.
Webvan was more ambitious and more expensive. An online grocery delivery service, it raised hundreds of millions, built automated warehouses across multiple cities before demand existed, and filed for bankruptcy in July 2001 having burned roughly $800 million. The idea was sound; online grocery is now a large business. It was attempted about fifteen years and one broadband generation too early.
Both were profitable ideas trapped in the wrong decade, funded by capital that assumed the funding would keep arriving.
The warnings that were ignored
The signals were visible. Companies were burning cash faster than they earned it and surviving purely on the next funding round. That is a solvable problem while capital is abundant and a fatal one the moment it is not, and everybody knew it.
Alan Greenspan asked in December 1996 how anyone could know when "irrational exuberance" had unduly escalated asset values. He was the chairman of the Federal Reserve, it was the most quoted phrase in finance for years, and the Nasdaq went up another 400% before he was proved right.
That interval is the reason bubbles are so hard to resist. The critics were correct and looked foolish for three consecutive years, while colleagues and neighbors got rich on paper. Plenty of careful investors held out, watched, and finally capitulated near the top — buying at the highest prices with the most conviction, which is what capitulation always looks like.
March 2000, and what came after
The Nasdaq peaked at 5,132 intraday on 10 March 2000. No single event broke it. Several things happened at once and sentiment turned.
The most concrete was monetary policy. The Federal Reserve raised rates six times between June 1999 and May 2000, taking the target from 4.75% to 6.5%. For companies whose survival depended on continuous access to cheap capital, that was existential rather than inconvenient. Funding dried up, weaker names started to fail, and each failure made investors ask the question they had been avoiding about the ones still standing.
Once the question was being asked, the mechanics reversed. The herd behavior that had lifted prices worked identically on the way down, and the companies that needed new capital could no longer raise it, so they failed, which confirmed the fear that had stopped the capital.
The Nasdaq bottomed in October 2002 at around 1,114, down roughly 78% from the peak. Thousands of companies disappeared. Even good businesses lost 80% to 90%.
One statistic captures the depth better than any other: the Nasdaq did not regain its March 2000 level until 2015, fifteen years later, and that is before adjusting for inflation. Anyone who bought the index at the top and simply held it waited half a working career to break even in nominal terms.
The survivors, and what made them different
Amazon is the case everyone cites. Its shares fell from over $100 to under $6 — a decline of roughly 95% — and serious people wrote it off. What Amazon had that most did not was a real business underneath the story: actual customers, actual logistics, and a founder who spent the boom building infrastructure rather than buying advertising. It survived because it was solvent when the money stopped.
Priceline, now Booking Holdings, fell even harder, losing over 99% of its value, and went on to become one of the best-performing stocks of the following two decades. eBay came through with an actual profit, which in 1999 had made it look boring.
The lesson is not "buy the crash." Survivorship is obvious only afterwards, and for every Amazon there were dozens of companies that looked similar and went to zero. The lesson is what separated them at the time: a business that generated cash, or credibly would, without needing the market to stay enthusiastic.
What it actually teaches
A correct thesis does not justify any price. The internet was the biggest commercial development in decades and buying it at the top still lost 78%. The quality of the idea and the price of the asset are two separate questions, and only one of them determines your return.
When the metrics change, be suspicious of the metrics. Eyeballs and mindshare were invented because earnings gave the wrong answer. Any time a market starts measuring value with something that has no path to cash, the measurement is doing marketing work.
"This time is different" is a signal, not an argument. Sometimes the technology genuinely is different. The rules for valuing cash flows are not.
Timing a bubble is close to impossible. Greenspan was three years early. Short sellers who identified the bubble correctly in 1998 were carried out before it broke. Recognizing that prices are unhinged says nothing about when they will stop being unhinged, and being early is indistinguishable from being wrong while it is happening.
The unwind is faster than the build. Five years up, thirty-one months down. This asymmetry is a permanent feature of markets and it is why the same crowd dynamics that create a bubble make the exit narrow. The same pattern appears in a short squeeze run in reverse.
The pattern keeps returning
The dot-com bubble stays relevant because its causes were psychological, and psychology does not get patched. Every genuinely transformative technology since has produced a version of the same sequence: an accurate long-term thesis, capital arriving faster than businesses can absorb it, valuations detached from any cash flow, new metrics invented to justify them, and a rush for the exit when funding conditions change.
That does not mean every new technology is a bubble. It means the questions are always the same ones. What is the business model? Does it produce cash, or is there a plausible path to it? What am I paying relative to that, and what has to go right for the price to make sense?
When those questions get answered with "you don't understand, the old rules don't apply," the answer is worth more than the question. Bubbles do not require anyone to be stupid. They require a good story, cheap money, and enough time for careful people to give up on being careful.
Markets crash for structural reasons too, and those look completely different — Black Monday in 1987 was a mechanical accident in a market that was not especially overvalued. Bubbles are the slower and more expensive failure, because they take years to inflate and take the real economy with them when they deflate.
