A recurring worry among investors right now is what happens if the enormous capital spending on AI infrastructure — data centers, chips, power — turns out to be unsustainable, and the cost of building and running AI systems outpaces the revenue it generates. History doesn't repeat exactly, but the dot-com era (roughly 1998-2002) is the closest real precedent we have for exactly this kind of dynamic, and it's genuinely instructive.
What actually happened in the dot-com era
The internet was a real, transformative technology — that part turned out to be entirely true. But in the late 1990s, capital poured into internet companies far faster than those companies could generate matching revenue or profit, based on the belief that growth and market share mattered more than near-term profitability. When financing conditions tightened and investor patience ran out around 2000-2001, companies that were burning cash without a path to profitability failed rapidly — the Nasdaq fell around 78% from its March 2000 peak to its 2002 low. Amazon's stock fell over 90% in that period.
The critical detail: the technology being right and the investment being overpriced are two separate facts, and both were true at once. The internet transformed the global economy exactly as its most bullish proponents predicted — and the vast majority of money invested in internet stocks during the bubble was still lost, because valuations had detached from any realistic near-term earnings.
Why AI spending has real similarities
Massive capital expenditure on AI infrastructure (chips, data centers, power generation) is being financed heavily on the assumption that AI-driven revenue will eventually justify it. If the actual usage and monetization of AI products grows more slowly than the infrastructure buildout, or if a cheaper technical approach reduces how much compute is actually needed, the companies most exposed to that spending — infrastructure providers, chipmakers, and AI-focused companies trading on future-growth expectations rather than current profits — would be the most vulnerable to a sharp repricing, similar to how internet infrastructure and "pure-play" dot-com companies were hit hardest in 2000-2002, while more established, profitable technology companies survived and eventually thrived.
The important distinction: a bubble popping isn't the same as the technology failing
After the dot-com crash, the internet didn't go away — it became more important than ever, and companies that survived the shakeout (Amazon, Google, among others) went on to become some of the most valuable companies in the world. The lesson isn't "avoid the technology" — it's that being right about a technology's importance and being right about specific company valuations are different questions, and conflating them is how investors get hurt in every major tech cycle, not just this one.
What a similar AI shakeout would likely look like
If AI infrastructure spending does outpace realistic near-term returns, a plausible pattern (based on the dot-com precedent) would be: a sharp repricing concentrated in the most speculative, least-profitable AI-adjacent companies; heavier, more sustained pain for companies that took on debt to fund the buildout versus those funding it from existing profits; and continued, possibly accelerated progress in the underlying technology itself, even as the stock prices of many companies tied to it fall substantially. The companies with real, profitable AI products and diversified revenue would likely be far more resilient than pure infrastructure or speculative plays.
What to actually do about this risk
- Avoid concentrating heavily in single AI-focused stocks or thematic AI funds — if this plays out anything like the dot-com era, the dispersion between "survived and thrived" and "wiped out" companies within the same theme was enormous.
- Remember that broad index funds already hold your AI exposure — most major indices are increasingly weighted toward large technology companies, so a diversified investor already has meaningful AI-linked exposure without needing a separate concentrated bet.
- Don't chase valuations that only make sense under best-case assumptions — a useful gut check from the dot-com era is asking whether a company's current valuation requires near-flawless execution and enormous future revenue growth to be justified, or whether it's reasonable under more conservative assumptions too.
- Stay invested through the underlying technology cycle even if specific stocks reprice sharply — a diversified, long-term investor who stayed in the market through the dot-com crash and its aftermath still captured the internet's genuine long-term economic impact; it was concentrated, leveraged bets on unprofitable dot-com names specifically that were wiped out.
Keep your plan diversified, not thematic
Our Investment Goal Calculator is designed around broad, diversified return assumptions rather than a single sector bet — the safer way to stay exposed to a genuinely important technology trend without betting your whole plan on which specific companies win the shakeout.