The similarities are impressive. Like internet companies 20 years ago, AI companies are attracting large investments today based on transformational potential rather than current profitability. Global Corporate AI investment reached $252.3 billion in 2024, according to a Stanford University survey, which has grown 13 times since 2014. Meanwhile, America's largest tech companies (Amazon, Google, Meta, and Microsoft) have pledged to spend a record $32 billion for AI infrastructure this year alone.
Even Openai CEO Sam Altman, who is valued at around $500 billion despite launching ChatGpt just two years ago, has admitted this similarity. ”
Are we at a stage where the entire investor is overly excited about AI? My opinion is yes,” Altman said in August. “Is AI the most important thing that should happen for a very long time? My opinion is yes too.”
But what lessons do the rupture of the dot-com bubble in March 2000 provide for today's AI boom? Take a walk along the memory lane. Also, if you haven't been born yet, have a plain-me-y history.
The perfect storm of 2000
The DOT-COM crash is not caused by a single event, but rather a convergence of factors that revealed fundamental weaknesses in the technological economy of the late 1990s. The first critical hit came from the Federal Reserve, which raised interest rates multiple times throughout 1999 and 2000. The federal funding rate rose from around 4.7% in early 1999 to 6.5% by May 2000, allowing investors to earn higher returns from safer bonds, making speculative investment unattractive.
The second catalyst was a wider economic recession that began in Japan in March 2000, causing fear in the global market and accelerated flights from dangerous assets. This higher 1-2 fee and global uncertainty have led investors to reevaluate the astronomical valuation of internet companies.
However, the underlying problem has deepened. Most Dot-Com companies had fundamentally flawed business models. Commerce One has reached a $21 billion valuation despite minimal revenue. TheGlobe.com was founded by two startup capital students at Cornell, with $15,000, and the stock price rose 606% on the first day of the transaction despite having no revenue beyond the venture's funds. Pets.com burned $300 million in just 268 days before declaring bankruptcy.
Infrastructure Overbuilding
Perhaps the most beneficial similarity of today's AI boom is the overinvestment of large infrastructure ahead of the DOT-COM crash. Telecommunications companies have built more than 80 million miles of fiber optic cables throughout the US, driven by Worldcom's much-increasing claim that internet traffic doubles every 100 days, above the actual annual doubling rate.
Companies such as Global Crossing, Level 3 and QWEST raced to build large networks to grasp unconstructed and anticipated demand. This resulted in a catastrophic excess. Even four years after the bubble burst, 85% to 95% of the fibers laid in the 1990s remained unused, earning the nickname “Dark Fiber.”
Corning, the world's largest fiber optic producer, won a stock price crash from about $100 in 2000 to about $1 by 2002. Ciena's revenues fell from $1.6 billion to $300 million, with 98% of its shares falling from its peak.
The similarities with today's AI infrastructure buildout are unmistakable. Meta CEO Mark Zuckerberg announced this year the AI data center's plan is “so big that it can cover a significant portion of Manhattan.” Backed by Openai, Softbank, Oracle and MGX, the Stargate project aims to develop a nationwide $500 billion network of AI data centers.
However, there are important differences. Unlike many DOT-COM companies that do not have revenue, the leading AI players generate substantial revenue. With an AI-focused Microsoft's Azure Cloud Service has increased 39% year-on-year to a run rate of $86 billion. Openai will project $20 billion in annual revenue by the end of the year, informationup from around $6 billion at the beginning of the year.
Big reality check
The DOT-COM crash eventually became a harsh reality. Most internet companies were unable to justify their valuations with actual business results. Companies were assessed based on website traffic and growth metrics, rather than traditional measures such as cash flow or profitability.
Today's AI companies are facing similar testing. Although AI investment has reached historic levels, the revenue gap remains large. According to technology writer Ed Zitron, it has invested around $560 billion in AI infrastructure over the past two years, bringing just $35 billion in AI revenue combined with AI-related revenue.
A recent MIT study found that 95% of AI pilot projects, despite over $40 billion in generated AI investments, are unable to produce meaningful results. This disconnect between investment and returns reflects the underlying problem that ultimately destroyed the dot-com bubble.
The question facing investors today is not whether AI will transform the economy. Most experts agree that it does. The question is whether current valuations and infrastructure investments can be justified by short-term returns, or whether much of today's AI infrastructure will not be used while many of today's AI infrastructure waits to keep up with supply, like the fiber optic cables of the 1990s. History shows that even transformative technologies cannot escape the gravity of economics. So it didn't happen as quickly as some of the early champions had promised while the internet was changing the world.
For this story, luck Generated AI was used to assist with initial drafts. The editors checked the accuracy of the information prior to publication.
