Everyone’s wondering if, and when, the AI bubble will pop. Here’s what went down 25 years ago that ultimately burst the dot-com boom

Yet, crucial differences exist. Unlike many dot-com companies that had no revenue, major AI players are generating substantial income. Microsoft’s Azure cloud service, heavily focused on AI, grew 39% year-over-year to an $86 billion run rate. OpenAI projects $20 billion in annualized revenue by the end of the year, according to The Information, up from around $6 billion at the start of the year.

The big reality check

The dot-com crash ultimately came down to a harsh reality: Most internet companies couldn’t justify their valuations with actual business results. Companies were valued based on website traffic and growth metrics rather than traditional measures like cash flow and profitability.

Today’s AI companies face a similar test. While AI investment has reached historic levels, the revenue gap remains substantial. According to tech writer Ed Zitron, Microsoft, Meta, Tesla, Amazon, and Google will have invested about $560 billion in AI infrastructure over the last two years, but have brought in just $35 billion in AI-related revenue combined.

A recent MIT study found that 95% of AI pilot projects fail to yield meaningful results, despite more than $40 billion in generative AI investment. This disconnect between investment and returns echoes the fundamental problem that ultimately doomed the dot-com bubble.

The question facing investors today isn’t whether AI will transform the economy—most experts agree it will. The question is whether current valuations and infrastructure investments can be justified by near-term returns, or whether, like the fiber-optic cables of the 1990s, much of today’s AI infrastructure will sit unused while the market awaits demand to catch up with supply. As history shows, even transformative technologies can’t escape the gravitational pull of economics—so while the internet did change the world, it didn’t happen as quickly as some of its early champions promised, and several of those people who got ahead of themselves were humbled in the process.

For this story, Fortune used generative AI to help with an initial draft. An editor verified the accuracy of the information before publishing.