Opinion: The AI Boom vs. the Dot-Com Era. What Are Their Main Differences?

As investors wonder whether the high valuations of companies benefiting from the AI revolution are justified, comparisons between the current situation and the dot-com bubble are becoming increasingly common. German Kaplun, co-founder of TMT Investments PLC, writes on his Facebook page about what distinguishes these two periods. Oninvest publishes his opinion in full and without changes.

***

Everyone is looking for a repeat of the dot-com bubble in AI.

The logic behind this comparison is simple: sky-high valuations and unrealistic expectations mean we’re headed for a repeat of 2000–2001.

But there are also differences.

Back then, the market was dominated by startups that were just getting started, fueled by expectations of future profits. Today, the AI race is led by the world’s largest companies, which generate tens and hundreds of billions of dollars and fund AI largely from their own cash flow, rather than just from endless funding rounds.

I tried to compare the leaders of two eras.

Cisco is a symbol of the Internet revolution. In 2000: revenue of ~$19 billion, profit of ~$2.7 billion, and a market capitalization at its peak of $555 billion. That’s ~29 times its annual revenue.

Nvidia today: revenue of $250 billion, $120 billion in net income, and a market capitalization of ~$5 trillion. That’s ~20 times its annual revenue—one-third lower than Cisco’s at the peak of the dot-com boom. At the same time, Nvidia’s annual profit today is 45 times greater than Cisco’s was in 2000. Cisco was overvalued for a dream. Nvidia is overvalued for a money-making machine that’s already up and running.

(In 2000, Google was still a small private company, so it would be more accurate to use Yahoo—the biggest star of the internet at that time.)

Yahoo in 2000: revenue of ~$1.1 billion, profit of ~$70 million, and a peak market capitalization of ~$125 billion in January 2000. That’s ~115 times its annual revenue.

OpenAI today: ~$25 billion in revenue (annualized run rate) at a valuation of ~$850 billion — ~34 times annual revenue. Expensive? Absolutely. But it’s still more than three times cheaper than Yahoo was in 2000—with a business on an incomparably larger scale.

Exodus Communications—one of the key players in the dot-com bubble—built data centers for the Internet of the future. In 2000, its revenue was approximately $800 million, and at its peak, its market capitalization stood at approximately $32 billion—about 40 times its annual revenue. But a year and a half later, following the dot-com crash, the company went bankrupt.

CoreWeave is already building infrastructure specifically for AI. Its revenue target for 2026 is $12–13 billion (year-end run rate of $18–19 billion), with a contracted backlog of approximately $100 billion. Its market capitalization is about $35 billion—that is, roughly three times its annual revenue. Not 40, like Exodus. Three.

And the key difference isn’t even in the multiplier—it’s in the clients. Exodus’s clients were internet startups that vanished along with the bubble. CoreWeave’s clients include Microsoft, OpenAI, Meta, and Nvidia—companies with multibillion-dollar cash flows and long-term contracts.

Does all this prove that the current valuations are fair and that there is no bubble? No, it does not.

And the very concept of a bubble is largely subjective: it all depends on your time horizon and your willingness to ride out periods when market sentiment shifts every couple of months. For a short-term trader, the bubble could burst as early as tomorrow. For a long-term investor, it may turn out to be just another stop along the path of years of growth—or decline... depending on choosing the right investment and a bit of luck.

This article was AI-translated and verified by a human editor