What Was the Dot-Com Bubble?

On March 10, 2000, the NASDAQ Composite closed at 5,048. That number, at the time, felt like evidence: evidence that the internet was remaking commerce, that the old rules about profits and revenue had expired, that anyone still asking whether a company could make money was simply behind. Within eighteen months, the index had surrendered nearly 70% of that peak.² By October 2002, the loss reached 78%.² The wealth that vanished was not abstract. Federal Reserve household balance-sheet data shows that Americans lost roughly $5 trillion in equity holdings across that period.⁶

This is the story of how it built, why it broke, and what it still teaches anyone who has ever watched a crowd sprint toward the same exit.

The Setup: A Genuine Revolution Meets Unlimited Money

The excitement was real. Between 1993 and 1995, the World Wide Web moved from university labs into living rooms. For the first time, a business could reach a national customer without a storefront, a print ad, or a television spot. Amazon proved you could sell books this way. eBay proved you could run an auction. The core observation, that the internet would reshape commerce and information, was not delusional.

What happened next was the leap. If the internet was genuinely transformative, investors reasoned, then any company with ".com" in its name would be worth billions, regardless of what it actually sold or whether it could ever cover its costs. The leap felt logical in the moment. It was not.

Perhaps the earliest public warning came from Federal Reserve Chairman Alan Greenspan, who in December 1996 asked, in a speech to the American Enterprise Institute, "how do we know when irrational exuberance has unduly escalated asset values?"¹ The market dipped briefly the morning after, then resumed climbing. Irrational exuberance, it turned out, had at least three more years of running room.

The Frenzy (1996 to 2000)

Let's walk through what the money actually looked like. Venture capital investment in internet companies reached roughly $100 billion across 1999 and 2000 combined.⁴ The logic driving those checks was straightforward in outline and catastrophic in practice: acquire users first, monetize later. Revenue was optional. Profit was, at best, a distant conversation.

Price-to-earnings ratios tell part of the story. Greenspan's retrospective analysis, delivered at Jackson Hole in August 2002, noted that the price-to-earnings ratio for the broader market rose from roughly 15 in 1995 to nearly 30 by 2000.² A P/E ratio (price-to-earnings, the amount investors pay per dollar of annual profit) of 30 is elevated but not unprecedented for high-growth companies. The problem was that many of the most celebrated internet companies had no earnings at all, making the ratio literally incalculable.

Consider a few representative cases. Pets.com, an online pet-supply retailer, reached a valuation near $300 million at its 2000 IPO. Its business model was, at its core, a policy of losing money on every sale by absorbing shipping costs that exceeded what customers paid. The company burned through $82 million in eighteen months and shut down in November 2000. Webvan, an online grocery delivery service, raised $375 million at IPO and eventually burned through more than $800 million building out warehouse infrastructure, never approaching profitability before its 2001 collapse. Priceline.com, which had a genuinely novel idea in name-your-price travel booking, reached a $13 billion market cap in 1999 despite having roughly $10 million in annual revenue, a price-to-sales ratio of 1,300-to-1.³

What those numbers mean is worth stating plainly: investors were pricing in more than a thousand years of that company's current revenue just to break even on their bet.

The Retail Investor and the Exit

Now shift to who was buying. The late 1990s saw a wave of retail participation in equity markets, accelerated by the spread of online brokerage accounts and financial media that treated day trading as a viable career path. The conviction driving it was sincere: tech stocks only went up, traditional valuation metrics were relics of the pre-internet economy, and anyone who hesitated was leaving money on the table. The Survey of Consumer Finances captures how deeply equities had penetrated household balance sheets by the turn of the century.⁵

This is the psychology at the center of every speculative bubble. The economist Jeremy Siegel, writing about long-run equity returns, identifies a recurring dynamic: when prices rise consistently, buyers interpret rising prices as confirmation that prices will keep rising, and new buyers enter on the basis of that interpretation rather than on any assessment of the underlying asset.³ The price increase becomes self-sustaining, right up until it isn't.

By spring of 2000, earnings reports began to disappoint. Cash burn at high-profile companies became public. Investors who had been buyers began to reassess. The exit was not orderly.

The Damage

The NASDAQ fell 34% between March and May 2000 alone. Many individual internet stocks fell further and faster: companies that had risen 500% or 1,000% in two years gave back those gains in months. The aggregate effect on household wealth was severe.⁶ GDP growth, which the Bureau of Economic Analysis had tracked at a robust pace through the late 1990s, softened sharply in 2001 as business investment pulled back and unemployment climbed.⁷ Venture capital effectively shut off. The phrase "internet company" became a liability rather than an asset; legitimate technology businesses deliberately distanced themselves from the label.

By 2001 and 2002, the Federal Reserve was engaged in aggressive rate cutting to limit the economic fallout, a response that would carry its own long-term consequences.²

Why Some Companies Survived

What separated Amazon, eBay, and the company that would become Google from the wreckage is worth examining carefully, because the answer was not that they were "better internet companies." The answer was unit economics.

Amazon lost money consistently through the 1990s, which made it look, in superficial headlines, like Webvan. The difference was that Amazon was losing money on fulfillment investment while its per-transaction margin was positive and improving. There was a credible path from loss to profit that did not require the laws of commerce to change. Amazon reached profitability in 2003 and has never looked back. Webvan's per-delivery economics were inverted: the more it delivered, the more it lost, and nothing about scale fixed that.

eBay never had the problem at all. It ran a marketplace, not a warehouse, which meant it collected fees without carrying inventory. By the time the bubble burst, eBay was already profitable, and the crash left its business model intact.

Google, founded in 1998, did not go public until August 2004, after the wreckage had cleared. Its search-advertising model tied revenue directly to user intent, a structure that proved both scalable and durable. The delay spared it from the euphoric overvaluation that doomed so many contemporaries.

What 2021 Proved the Lesson Was Not Learned

Overall, the dot-com cycle is not a story about technology being overestimated. The internet became precisely as important as the optimists believed. The story is about what investors priced into companies in the absence of any discipline around when, and whether, those companies would generate returns.

The same dynamic materialized in 2020 and 2021. Pandemic conditions accelerated digital adoption, and a new generation of unprofitable technology companies, from electric-vehicle startups to consumer fintech platforms, reached billion-dollar valuations on narratives about future market share. When the Federal Reserve began raising interest rates in 2022, the discount rate applied to distant future profits rose sharply, and companies with no near-term earnings saw their valuations compress by 70%, 80%, even 90% in some cases. The shape of the chart was familiar.

History does not repeat exactly. The institutions involved, the sectors, the specific companies all change. What persists is the pattern: a genuine technological or economic shift, a leap from "this is real" to "everything with this label is worth any price," a period in which the crowd's momentum substitutes for individual judgment, and a correction that arrives faster than most participants expect and costs more than they had planned.

The internet did not disappear in 2002. The question was never whether the technology mattered. The question was whether you had paid a fair price for the share of it you owned. That question never goes away.