The similarities between today's AI firms and the internet companies of two decades ago are striking. Just as the early internet companies attracted massive investments based on their transformative potential rather than current profitability, the AI sector is experiencing a similar surge of interest. Research from Stanford University indicates that global corporate AI investment reached an astounding $252.3 billion in 2024, marking a thirteenfold increase since 2014. Major players in the tech industry, including Amazon, Google, Meta, and Microsoft, have committed to spending a record $320 billion on capital expenditures this year, much of which is directed toward AI infrastructure.
Even OpenAI CEO Sam Altman, whose company has achieved a valuation of approximately $500 billion just two years after launching ChatGPT, recognizes these parallels. In August, Altman remarked, “Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes.” However, this raises an important question: What caused the dot-com bubble to burst in March 2000, and what lessons can we glean for today’s AI boom?
The dot-com crash was not triggered by a single event, but rather a combination of factors that exposed significant weaknesses in the late 1990s tech economy. One of the first critical blows came from the Federal Reserve, which raised interest rates multiple times throughout 1999 and 2000. The federal funds rate climbed from around 4.7% in early 1999 to 6.5% by May 2000, making speculative investments less appealing as investors could secure better returns from safer bonds.
A broader economic recession that began in Japan in March 2000 further fueled global market fears and accelerated the flight from risky assets. This combination of higher interest rates and global uncertainty led investors to reassess the astronomical valuations of internet companies. However, the underlying issue was even more profound: many dot-com companies operated with fundamentally flawed business models. For instance, Commerce One reached a $21 billion valuation despite generating minimal revenue, while TheGlobe.com, founded by two Cornell students with just $15,000 in startup capital, saw its stock price skyrocket by 606% on its first day of trading, despite having no revenue beyond venture funding.
One of the most instructive parallels between the dot-com era and today’s AI boom lies in the massive infrastructure overinvestment that preceded the dot-com crash. Telecommunications companies laid over 80 million miles of fiber optic cables across the U.S., driven by inflated claims that internet traffic was doubling every 100 days—a far cry from the actual annual doubling rate. Companies like Global Crossing, Level 3, and Qwest rushed to build extensive networks to capture anticipated demand that ultimately never materialized, resulting in catastrophic overcapacity.
Years after the bubble burst, 85% to 95% of the fiber laid in the 1990s remained unused, earning the nickname “dark fiber.” For example, Corning, the largest optical-fiber producer, saw its stock plummet from nearly $100 in 2000 to about $1 by 2002. Similarly, Ciena experienced a dramatic decline in revenue, falling from $1.6 billion to $300 million almost overnight, with its stock plummeting by 98% from its peak. Today, the parallels to the current AI infrastructure buildout are unmistakable, as major players like Meta and OpenAI are making significant investments in AI data centers.
Despite the massive investments in AI infrastructure, the reality check for today’s AI companies mirrors that of the dot-com era. The dot-com crash ultimately boiled down to the fact that most internet companies could not justify their valuations with actual business results. Companies were often valued based on website traffic and growth metrics rather than traditional measures like cash flow and profitability. Today, AI companies face a similar challenge.
Although AI investment has reached historic levels, the revenue gap remains substantial. According to tech writer Ed Zitron, Microsoft, Meta, Tesla, Amazon, and Google are projected to invest around $560 billion in AI infrastructure over the past two years but have collectively generated only $35 billion in AI-related revenue. Additionally, a recent MIT study found that 95% of AI pilot projects fail to yield meaningful results, despite over $40 billion in generative AI investment. This disconnect between investment and returns echoes the fundamental issue that ultimately led to the dot-com bubble's demise.
The pressing question for investors today isn’t whether AI will transform the economy—most experts concur that it will. Instead, the inquiry focuses on whether current valuations and infrastructure investments can be justified by near-term returns, or if, like the fiber-optic cables of the 1990s, much of today’s AI infrastructure will remain underutilized as the market awaits demand to align with supply. Historical precedents show that even the most transformative technologies cannot escape the laws of economics. While the internet did indeed change the world, the process was far slower than its early advocates predicted, and many who overestimated their predictions faced significant repercussions.
