Everyone wonders if, and when, the Ai bubble will appear. Here is what happened 25 years ago which finally erupted the boom of the commun points

The similarities are striking. Like Internet companies two decades ago, AI companies today attract massive investments based on potential transformer rather than current profitability. The global investment on corporate AI reached $ 252.3 billion in 2024, according to research from the University of Stanford, the sector being thirteen times since 2014. Meanwhile, the largest technological companies in America – Amazon, Google, Meta and Microsoft – were committed to spending 320 billion in capital this year, a large part of the infrastructure.

Even the CEO of Openai, Sam Altman, whose company is estimated at around $ 500 billion despite the launch of Chatgpt only two years ago, recognizes parallels. “”

Are we in a phase where investors as a whole are overexcrited about AI? My opinion is yes, “said Altman in August.” Is AI the most important thing to happen for a very long time? My opinion is also yes.

But what really had the DOT-COM bubble exploded in March 2000, and what lessons does it offer it for today’s AI boom? Let’s take a walk in the past – or, if you were not yet born, a simple story.

The perfect storm of 2000

The Crash of the Comme Points was not triggered by a single event, but rather by a convergence of factors which exhibit fundamental weaknesses in the technological economy of the late 1990s. The first critical hit came from the Federal Reserve, which increased interest rates several times in 1999 and 2000. The rate of federal funds rose from around 4.7% at the start of 1999 to 6.5% 2000, which makes speculative investments less attractive because investors could obtain higher yields of safer bonds.

The second catalyst was a broader economic recession that started in Japan in March 2000, triggering fears of the world market and accelerating the leak of risky assets. This one-two highest rates of higher rates and global uncertainty led investors to reassess astronomical evaluations of Internet companies.

But the underlying problem was much deeper: most Dot-Com companies had fundamentally imperfect commercial models. Trade has reached an assessment of $ 21 billion despite minimal turnover. Theglobe.com, founded by two Cornell students with $ 15,000 in start-up capital, saw its grant price jump 606% on its first day of negotiation at $ 63.50, despite any income beyond the financing of the company. Pets.com burned $ 300 million in just 268 days before declaring bankruptcy.

Excessive infrastructure

Perhaps the most instructive parallel for today’s AI boom lies in the overinvestment of massive infrastructure which preceded the dot-com accident. Telecommunications companies have posed more than 80 million kilometers of fiber optic cables in the United States, driven by the claim of Worldcom, WorldCom’s fleet that Internet traffic was doubled every 100 days – further from the real annual doubling rate.

Companies like Global Crossing, Level 3 and Qwest ran to build massive networks to enter the planned demand that has never materialized. The result was catastrophic overcapacity. Even four years after the bubble bursting, 85% to 95% of the fibers posed in the 1990s remained unused, winning the nickname “black fiber”.

CORNING, the largest producer of optical fibers in the world, saw its actions crash from almost $ 100 in 2000 to around $ 1 in 2002. CIENA’s revenues increased from 1.6 billion to 300 million dollars almost overnight, its actions plunging 98% compared to its peak.

The parallels with the construction of today’s AI infrastructure are undoubtedly. The meta-PDG Mark Zuckerberg announced this year plans for an AI data center “so large that it could cover a significant part of Manhattan”. The Stargate project, supported by Openai, SoftBank, Oracle and MGX, aims to develop a national network of dollar centers of $ 500 billion.

However, crucial differences exist. Unlike many Dot-Com companies that had no income, the main AI players generate substantial income. Microsoft’s Cloud Azure service, which is highly focused on AI, increased by 39% on the other at an execution rate of $ 86 billion. Openai projects $ 20 billion in annualized income by the end of the year, according to InformationAgainst about $ 6 billion at the start of the year.

The great verification of reality

The Crash of the Comme Points finally reached a harsh reality: most Internet companies could not justify their assessments with real trade results. Companies were valued according to website traffic and growth measures rather than traditional measures such as cash flows and profitability.

Today’s IA companies are faced with a similar test. While investment in AI has reached historical levels, the income gap remains substantial. According to technological writer Ed Zitron, Microsoft, Meta, Tesla, Amazon and Google will have invested around $ 560 billion in AI infrastructure in the past two years, but have reported only $ 35 billion in revenue linked to combined AI.

A recent MIT study revealed that 95% of AI pilot projects do not give significant results, despite more than $ 40 billion in generative AI investment. This disconnection between investment and yields echoes the fundamental problem which finally condemned the Dot-Com bubble.

The question facing investors today is not if AI will transform the economy – most experts are suitable. The question is whether current assessments and investments in infrastructure can be justified by short -term yields, or if, like fiber optic cables of the 1990s, a large part of today’s IA infrastructure will remain unused while the market awaits the demand to make up for supply. As history shows, even transformative technologies cannot escape the gravitational attraction of the economy – so even if the Internet has changed the world, this did not occur as quickly as some of its first champions promised, and several of these people who took ahead were humiliated in the process.

For this story, Fortune Used a generative AI to help an initial project. An editor checked the accuracy of the information before the publication.

Global Forum fortune returns on October 26 to 27, 2025 in Riyadh. CEOs and world leaders will meet for a dynamic event only invitation that shapes the future of business. Request an invitation.