Why did the dot-com bubble burst, and does the AI bubble look the same?

What caused the dot-com bubble to burst in 2000?

The dot-com bubble burst because the cash that fed it ran out before the revenue arrived. Many internet companies sold growth, and investors paid for growth with new share sales. Each company spent on advertising and servers ahead of income, so the business model needed fresh money every few months. When fresh money slowed, prices adjusted to what the companies actually earned.

Three conditions sat behind that chain: easy funding, shares priced on future users instead of present profit, and a crowd of buyers who expected to sell higher. Each condition is measurable, which makes the burst easier to read afterwards than during the boom. You can check the same three conditions in any market, including AI today.

Which events triggered the dot-com bubble burst in March 2000?

The Nasdaq Composite peaked on 10 March 2000 at 5,048.62 (Wikipedia, Dot-com bubble). On 20 March 2000 Barron’s ran a cover story titled Burning Up, which warned that internet companies were running out of cash fast. The article gave buyers a shared reason to ask where the money came from. A single headline rarely ends a bubble, yet it can turn a private doubt into a public question.

On 3 April 2000 a US court found Microsoft guilty of monopolization, Microsoft shares fell 15% in one day, and the Nasdaq dropped 350 points, or 8% (same source). The sequence shows how a bubble ends: a doubt about cash, then a shock to a leading company, then sellers outnumber buyers. The page on when the dot-com bubble burst lists these dates and the later ones in order.

How far did prices fall after the dot-com bubble burst?

From the March 2000 peak, the Nasdaq Composite lost 78% of its value by October 2002 (Wikipedia, Dot-com bubble). Most internet stocks had already fallen 75% from their highs by November 2000, which wiped out $1.755 trillion of value. A fall of that size takes years, not weeks, and it comes in waves with rebounds in between.

The guide to the dot-com crash and the Nasdaq follows the index step by step. Prices fell faster than the technology faded, and that gap is what makes a bubble costly for holders: the internet kept spreading while valuations reset. A reader who wants the full path of the index will find it there.

Which companies survived the dot-com bubble burst, and why?

The same Wikipedia article names Amazon and eBay as companies that gained market share and came to dominate their fields, and it reports that 48% of dot-com companies survived through 2004, at lower valuations. Cisco lost 80% of its stock value. Survival and a high share price are two different things: a firm can live on while its shares fall.

Survivors share habits that you can check in any cycle: customers who pay, costs that fall as volume grows, and a balance sheet that lasts through the low point. The guide on dot-com survivors examines Amazon, eBay, and Cisco in turn. Applied to AI, the habits become a screen for which firms hold up when funding slows.

Is Cisco in 2000 the right comparison for the AI bubble?

Cisco sold the equipment that the internet ran on, so its sales rose with everyone else’s spending. When customers cut spending, the supplier felt it first. That makes Cisco the usual reference for companies that sell the hardware of a boom, and its stock lost 80% (Wikipedia, Dot-com bubble) while the internet kept growing.

The page Cisco in 2000 vs Nvidia today sets the two cases side by side through criteria: who the customers are, how they fund purchases, and how fast orders can change. It offers a method and leaves forecasts to others, since this page is general information, not investment advice.

Does the AI bubble look the same as the dot-com bubble?

Two features match. Valuations rose fast, with the study recording a ×4 rise in 18 months, and spending is large, about $200B a year (EQOS Asia AI study, April 2026). One feature differs: many AI buyers are large companies with cash flow. The study measures the enterprise market at φ(S) = 0.337, in the fragile zone, where a shock can tip the system.

The study also adds a signal about results: 95% of AI projects fail in production (MIT 2025, cited by the study). A comparison is useful when it names what repeats and what changes. Here the funding chain repeats, the buyer base looks stronger, and the project results look weaker.

What can a decision maker do about the AI bubble now?

Use the dot-com chain as a checklist. Ask where the cash comes from, whether customers pay, and how many months of funding the business holds. Write the answers down each quarter, so your view follows evidence. Staging large commitments keeps options open whichever way the market settles, and the approach suits cautious and growth-minded plans alike.

The EQOS Asia method page explains how the φ(S) score is built, and the Λ simulator lets you test your own case in 10 minutes. A 30-minute pre-diagnostic applies the same scoring to a decision in Singapore, ASEAN, or Hong Kong. This page is general information, not investment advice.

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