AI bubble vs dot-com bubble: what is the same and what is different?

What do the AI bubble and the dot-com bubble share, and where do they differ?

Both bubbles surround a real technology that changed how companies work. In both, prices rose faster than earnings, funding came easily, and the story about the future carried the valuation. These shared traits explain why the dot-com bubble is the first comparison anyone makes. The comparison earns its place when you use it as a checklist: for each trait, ask whether the AI market shows it today, and write the answer down with a date.

The differences matter as much as the similarities. After the Telecommunications Act of 1996, telecom equipment companies invested over $500 billion, mostly financed with debt (Wikipedia, Dot-com bubble). The EQOS Asia AI study measures a different market, which spends about $200B a year and scores 0.337 on the φ(S) scale. Each difference changes how much weight the comparison deserves.

What was the dot-com bubble, and what does it teach about the AI bubble?

The dot-com bubble was a run of rising prices for internet companies that peaked on 10 March 2000, when the Nasdaq Composite reached 5,048.62 (Wikipedia, Dot-com bubble). The first lesson is scale: a whole index, and not a few shares, priced in a story. The guide on what the dot-com bubble was tells the full story, from the early listings to the last weeks of the rally.

The technology outlived the prices. Internet use kept its place in daily life when valuations fell, so the technology and the price level followed different paths. AI can show the same split. The EQOS Asia AI study projects productivity gains of +40% next to 95% of AI projects failing in production (MIT 2025), which is a real technology beside fragile results.

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

A bubble ends when the funding that carries it slows. The Nasdaq Composite fell 78% from its peak by October 2002 (Wikipedia, Dot-com bubble), which shows how wide the gap between price and earnings had grown. The guide on why the dot-com bubble burst separates the triggers, such as rising interest rates, missed earnings, and a loss of confidence, and ranks them by weight.

The AI version of that question is which trigger is visible today. Funding, earnings, and rules are the 3 to watch, and each can be observed every month. The same guide holds them against AI and shows where the match is close and where it breaks. Reading triggers beats guessing dates, because a trigger tells you what to do when it appears.

How long did the dot-com bubble last, and what does it say about AI?

Measured from the Nasdaq peak of March 2000 to the low of October 2002, the fall took about 31 months. The rise before it took years longer, which is why a bubble looks safe while it grows. The guide on how the dot-com bubble lasted lays out both phases and shows what each one looked like to people living through it.

For the AI bubble the length matters in a practical way. If a run can last years, the statement that a bubble exists tells you little about when to act. Decision makers therefore set rules in advance: the size of exposure, review dates, and triggers. The rules work whichever way the length turns out. The EQOS Asia method page explains how a score supports such rules.

How does the 2008 crisis compare with the AI bubble?

The 2008 crisis was a credit bubble, while the dot-com bubble was an equity bubble. The difference matters because credit spreads losses through banks to the whole economy. The EQOS Asia 2008 back-test records house prices up 124% between 1997 and 2006 and a subprime share of 23.5% against 8% in 2003 (EQOS Asia 2008 back-test, 2026). The guide on the 2008 crisis vs the AI bubble sets those risks beside AI.

The question for AI is which channel could carry a loss to the wider economy: shares, bank loans, or company debt. Mapping the channel tells you how far a fall could travel. The same back-test scored the 2006 financial system at φ(S) = 0.234, in the critical zone, against 0.337 for enterprise AI today, in the fragile zone (EQOS Asia AI study, April 2026).

Where do crypto and stock market concentration fit in the AI bubble comparison?

Crypto is the recent mania most people remember. The EQOS Asia crypto analysis records the FTX collapse of November 2022, with $8 billion of customer funds lost, and Bitcoin swings of ±30% per month (EQOS Asia crypto study, 2026). The guide on crypto bubble vs AI bubble compares the two manias on funding, regulation, and who holds the risk.

Stock market concentration is a different lens: the question is how much of an index rests on a few AI names. The guide on the stock market bubble question explains how to measure that share and how to read it. Both lenses help you judge whether AI stress would stay inside one sector or spread across portfolios.

Where does the AI bubble sit in bubble history, from tulips to dot-com?

Bubbles repeat a pattern more than a calendar: a real novelty, easy money, a story that grows, then a reversal. The guide on bubbles through history puts the AI bubble in that line, with tulips and railways beside dot-com. Each case adds one test you can apply to AI: who funded it, what it produced, and what remained after it.

Use the history as a method, then add measurement. The EQOS Asia Λ simulator estimates φ(S) for your own case in 10 minutes, free and anonymous. For a decision on AI spending in Singapore, ASEAN, or Hong Kong, the pricing page lists the formats available. As of October 2026 the open question stays the same: how much of the AI price rests on earnings.

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