The AI bubble has surpassed the dot-coms: the scales are incomparable
The artificial intelligence market is experiencing explosive growth, which in several key parameters has already left behind the legendary dot-com bubble of the late 1990s. As an analyst, I closely monitor these processes, and the parallels between the two eras are indeed hard to ignore. However, the current AI frenzy is not just repeating history—it surpasses it in scale and depth of penetration into the financial system.
AI Infrastructure: 10 Times Larger Than the Internet
The key difference lies in the volume of capital investments. During the dot-com era (1999–2001), total infrastructure spending amounted to about $500 billion. Today, during the AI boom (2025–2026), this figure has soared to $5.8 trillion. This is not just growth—it is an entirely different order of magnitude.
A similar picture is observed in the primary offering market. The largest IPO during the dot-com era raised around $4 billion. In the current cycle, record holders are already operating with sums of $86 billion. The difference is more than 20 times.
Unprofitable Giants: A New Reality
Particular attention should be paid to the valuation of unprofitable companies. Twenty years ago, the largest "loss-maker" was worth about $100 billion. Today, its equivalent is valued at $965 billion. Moreover, many of these companies go public at the peak of the hype and then rapidly lose market capitalization. A telling example: SpaceX, after the largest IPO in market history, went into the red within a month.
Dangerous Signals: Customer Lending
Another alarming sign is financing schemes. Nvidia, like its predecessors Lucent and Nortel in the late 1990s, has launched a program to lend to its own customers. For those who remember history, this is a direct path to some of the largest bankruptcies in U.S. history. When a chip manufacturer starts financing the purchase of its own products, it is a classic sign of overheating.
Multi-Layered Structure: Risks Broader Than They Seem
The main danger of the current bubble is that it is not limited to stocks. Unlike the dot-com era, where speculation mainly involved the stock market, today's AI boom has absorbed private credit, project bonds, and even insurance company funds. All of this flows into data centers, creating a multi-layered debt structure.
Examples are already on the surface. Blue Owl froze investor withdrawals while simultaneously financing 80% of Meta's (banned in Russia) $250 billion campus. Oracle's rating was downgraded to nearly "junk"—half of its order book is tied to OpenAI. Anthropic has $90 billion in capacity lease obligations with zero profit before its IPO.
My conclusion as an analyst: AI technology is real and promising—just like the internet in 1999. But the issue is not the technology, but the price. Do the companies justify the sums investors are paying for them? Public reports will soon provide the answer. For now, I recommend caution: the history of the dot-com era teaches that even great technologies can lead to catastrophic losses for those who enter an overheated market without regard for fundamental indicators.