The dot-com crash and the Nasdaq: how far prices fell, and what it means for the AI bubble

How far did the Nasdaq fall in the dot-com bubble crash?

The Nasdaq Composite peaked at 5,048.62 on 10 March 2000 and had lost 78% of its value by October 2002 (Wikipedia, Dot-com bubble). That is the headline number of the dot-com crash. Put differently, 100 invested at the peak stood at 22 at the low, so a holder of the whole index lost more than three quarters of the money in about 2 years and 7 months.

A fall of that size also changes behaviour. Buyers who held through the descent watched each rebound fail, employees saw option value vanish, and founders lost room to raise new money. The Nasdaq carried the technology names that had risen fastest, so the crash concentrated in the segment where expectations had run furthest ahead of earnings.

How big a rebound do you need after a 78% dot-com loss?

Percentages fall and rise on different bases. A fall of 25% needs a gain of 33% to recover. A fall of 50% needs 100%. A fall of 75% needs 300%. A fall of 78%, the dot-com figure, needs a gain of about 355%. The deeper the fall, the more the recovery depends on years of new growth, not on a quick rebound.

This arithmetic is why a crash can last long after the headline day. Capital sits waiting, plans get cut, and funding for new projects tightens while prices climb back. For a business that depends on outside money, the recovery period matters as much as the fall, because the funding gap lasts as long as the climb.

How fast did prices drop during the dot-com crash?

The drop came in waves, not in a straight line. On 3 April 2000 the Nasdaq lost 8% in one session, about 350 points, after the Microsoft ruling, and Microsoft shares fell 15% that day (Wikipedia, Dot-com bubble). By November 2000, most internet stocks stood 75% below their highs. Rebounds appeared between the waves and tempted buyers back in.

The speed differed by stage. The first leg was fast and driven by the end of easy funding. The second leg was slower and driven by results: companies reported losses, laid off staff, and closed. A reader who tracks a market today can separate the two stages by asking whether prices move on funding news or on earnings news.

Which dot-com stocks fell less than the index, and which fell more?

The spread between stocks was wide. Cisco lost 80% of its stock value, while Amazon and eBay gained market share and came to dominate their fields (Wikipedia, Dot-com bubble). Pets.com, Webvan, and Boo.com closed. The index average of 78% hides companies that fell to zero and companies that later grew far beyond their old highs.

That spread offers the main lesson for selection. A crash punishes the weakest balance sheets first and rewards firms with paying customers. Looking at survivors one by one, as the guide to dot-com survivors does, turns a market-wide number into questions about cash, customers, and costs that you can ask of any company.

How much value did the dot-com bubble crash wipe out?

Wikipedia reports that the fall of most internet stocks by November 2000 wiped out $1.755 trillion of value. That figure measures paper value, which means market prices at the time, and it shows the scale at which a crash redistributes wealth. Many buyers lost savings, many employees lost options, and many founders lost control of companies they had financed with shares.

Compare it with the EQOS Asia AI study, which cites Bain & Company on $2.3 trillion of digital-transformation projects failing every year across the world. The two figures measure different things, a one-time price reset and a yearly project loss, yet both show the order of magnitude at which technology waves can destroy value.

What would a dot-com style crash mean for the AI bubble?

The EQOS Asia AI study records AI valuations up ×4 in 18 months and about $200B a year invested (EQOS Asia AI study, April 2026). The arithmetic is direct: a rise of ×4 reverses completely with a fall of 75%, because 4 times 25% equals 1. A fall of 78% would take valuations below their starting point.

This calculation describes a scenario, not a forecast. As of October 2026 a crash has not been confirmed, and the study scores the enterprise AI market at 0.337 in the fragile zone, where a shock can tip the system. The useful question is how much of your exposure would remain viable if prices returned to the level before the rise.

How can you size AI bubble exposure with dot-com crash numbers?

Run three scenarios: a fall of 25%, a fall of 50%, and a fall of 78%. For each, write the loss in money and the gain needed to recover. Then stage large commitments so that the middle scenario leaves the business intact. The EQOS Asia method page describes how such scenarios feed a repeatable score.

The Λ simulator returns a φ(S) estimate in 10 minutes, and a 30-minute pre-diagnostic applies the scoring to a decision in Singapore, ASEAN, or Hong Kong. This page is general information, not investment advice, and the dot-com figures describe the past rather than a prediction for AI.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile