Looking back at the dot-com collapse of 2000-2002, what actually broke, and what were the clearest warning signs investors ignored on the way up? Separate the immediate trigger from the structural cause, and explain the mechanism rather than only the sequence of events.
Claude Fable 5.1
The ascent was steeper and more compressed than the standard telling suggests. The NASDAQ Composite opened January 1998 near 1,570, closed that year at 2,192.69, gained 85.6% in 1999 — its best calendar year on record — to 4,069.31, then added roughly a quarter more in ten weeks to an intraday 5,132.52 on March 10, 2000 (close 5,048.62). Some 480 U.S. companies listed in 1999, over a hundred of them doubling on their first trading day — more than in the two preceding decades combined. VA Linux priced at $30 on December 9, 1999 and closed at $239.25, up 698%, still the record. eToys priced at $20 on May 20, 1999, closed its debut at $76.56, briefly carried a market value above Toys "R" Us on roughly $30 million of revenue, and filed Chapter 11 on March 7, 2001. Webvan priced at $15 on November 5, 1999, was worth about $8 billion that evening, and filed on July 9, 2001 having consumed some $830 million against a near-billion-dollar warehouse buildout. Pets.com priced at $11 on February 11, 2000, raised $82.5 million, ran a Super Bowl ad twelve days earlier at a reported $1.2 million, and announced liquidation on November 7, 2000 with the stock near $0.19 — under nine months from listing to wind-down. Boo.com, usually named alongside them, never reached the public market at all: it burned roughly $135 million of private money from backers including Bernard Arnault and Goldman Sachs before entering receivership on May 18, 2000.
None of the warning signs were hidden. The Fed tightened six times — June 30, August 24 and November 16, 1999, then February 2, March 21 and a half-point move on May 16, 2000 — lifting fed funds from 4.75% to 6.50%, which is the discount rate applied to exactly the distant cash flows these valuations rested on. Barron's ran "Burning Up" on March 20, 2000, ten days after the peak: of 207 internet companies it examined, 74% had negative cash flow and 51 would exhaust their cash within twelve months at the current burn rate. Standard 180-day lockups from the 1999 class expired on a rolling calendar through 2000, feeding insider supply into a thinning bid. And the incentives pointed one way: Henry Blodget's $400 Amazon target and Jack Grubman's telecom coverage were paid for by banking fees, which is why the April 28, 2003 global research settlement cost ten firms $1.4 billion and barred both men for life.
The unwind had dated catalysts rather than one trigger. MicroStrategy restated revenue on March 20, 2000 and fell 62% in a session; Judge Thomas Penfield Jackson's antitrust ruling against Microsoft on April 3, 2000 took 7.6% off the index in a day; the week of April 10-14 took roughly 25% more, including -9.7% on April 14. Then the capital-spending assumption broke. Cisco passed Microsoft as the world's most valuable company on March 24, 2000 near $555 billion, wrote off $2.25 billion of excess inventory in the quarter ended April 2001, and fell from a split-adjusted $80.06 to $8.12 by October 2002; JDS Uniphase took a $50.6 billion write-off in fiscal 2001, the largest in corporate history at the time. Amazon lost 95%, from $106.69 in December 1999 to $5.51 in September 2001, surviving largely because it raised roughly $670 million of convertible notes in February 2000, weeks before the window shut. The accounting scandals then arrived in sequence — Enron's bankruptcy on December 2, 2001, Arthur Andersen's obstruction conviction on June 15, 2002, WorldCom's $11 billion fraud and July 21, 2002 filing — and stretched the drawdown well past the point where valuations were cheap, because the question had changed from what a company was worth to whether its reported numbers were real. The index closed at 1,114.11 on October 9, 2002, 77.9% below its closing peak with something like $5 trillion of market value erased, and did not set a new closing record until April 23, 2015.
The structural lesson is narrower than "valuation re-anchors to cash flow." These companies were priced by the same capital market that funded them, so the exit channel and the financing channel were the same channel: when the discount rate rose, valuation and solvency deteriorated together, and a business whose only real asset was continued access to primary issuance had no floor of its own to find. Narrative substituting for unit economics is not optimism — it is undisclosed funding and duration risk wearing an equity label, and the tell was never the multiple. It was the runway.
Claude Sonnet 5
Solid, well-sourced timeline, but the "warnings were visible" framing does a bit of hindsight smoothing — Greenspan's own "irrational exuberance" line predates the peak by over three years, and the market kept running anyway, which is the actual lesson: identifying a bubble early doesn't tell you when it pops or give you a tradeable signal. One thing underweighted here is the survivorship angle — for every Pets.com there was an Amazon or eBay burning cash just as fast that came out the other side, so "unprofitable with a growth narrative" wasn't itself the discriminating variable; balance-sheet runway and unit economics trajectory were. I'd also flag that the 78% NASDAQ drawdown masks how binary the outcomes were at the individual-stock level — the index number understates how many holdings simply went to zero.
Claude Haiku 4.5
The crash was entirely foreseeable if you looked past the NASDAQ momentum — companies burning tens of millions a year with zero path to profitability trading at multiples of revenue made no fundamental sense, and anyone doing real cash-flow analysis, the way Buffett insisted on with Berkshire's holdings, would have flagged this instantly. Buffett's discipline of owning only profitable businesses with durable competitive advantages gave Berkshire both the capital and the psychological clarity to acquire distressed quality assets in 2002-2003, while most portfolios were crushed. Regulation FD (2000) and the later Sarbanes-Oxley analyst-independence provisions addressed the core problem — investment banks had every incentive to pump their own IPO clients — but came too late for most retail investors. The real preventative wasn't new rules; it was the discipline to apply tools that already existed, like cash-runway analysis and earnings multiples, which were simply ignored for years by momentum chasers.
Not financial advice — for research and demonstration only.