In 1999, everyone knew the internet was going to change the world. They were right. They were also completely wrong about which companies would survive, what the timeline would look like, and how painful the crash would be.
Twenty-six years later, serious investors are asking whether we’re in the same movie again. Just with artificial intelligence replacing the internet as the story.
Here’s the honest comparison nobody wants to have.
What Happened in the Dot-Com Bubble
From January 1995 to March 2000, the Nasdaq Composite surged 572%, from 751 points to a peak of 5,048, driven almost entirely by internet stocks.
Investors poured money into any company with a “.com” in its name. Profits didn’t matter. Revenue barely mattered. The story mattered. The idea that the internet would transform every industry created a buying frenzy that pushed valuations to levels that had no basis in reality.
Then reality caught up. Growth expectations proved too optimistic, many unprofitable companies ran out of cash, and interest rates had been rising. The bubble burst spectacularly. Over the subsequent two years, the Nasdaq plunged nearly 80% from its peak, while the S&P 500 fell almost 50%. More than $5 trillion of market value evaporated, and scores of dot-com darlings went bankrupt.
The internet didn’t fail. Amazon survived. Google was born. The technology fulfilled every promise it made. But the investors who bought in 1999 at peak valuations waited over a decade to get their money back.
What’s Happening With AI Right Now
The parallels are hard to ignore.
Nvidia’s share price is up over 1,300% since late 2022. OpenAI is valued at $852 billion. Anthropic at $380 billion. Databricks at $134 billion.
Capital expenditure spending by the biggest tech companies is expected to rise to half a trillion dollars in 2026 alone. Microsoft, Google, Amazon, and Meta are collectively spending more on AI infrastructure this year than the GDP of many medium-sized countries.
The S&P 500 has topped 7,300 for the first time ever. The Nasdaq has capped a six-week advance. One investment chief described the current moment as “like the internet on steroids. And we’re still in the second or third inning.”
Sound familiar?
Warren Buffett’s preferred market valuation measure, the ratio of total stock market value to US GDP, reached a record high above 140% during the dot-com boom, right before the crash. Today it sits above 210%. “If the ratio approaches 200%, you are playing with fire,” Buffett wrote in 2001.
It’s at 210% today. He said 200% was playing with fire.
Torsten Slok, chief economist at Apollo Global Management, recently argued that the stock market currently overvalues a handful of tech giants, including Nvidia and Microsoft, even more than it overvalued early internet companies on the eve of the 2000 dot-com crash. He says 12-month forward price-to-earnings ratios now exceed the peak of the dot-com bubble.
The Key Differences — Why This Might Not End the Same Way
Here’s where the comparison gets more complicated. Because there are meaningful differences between 1999 and today and they matter.
The biggest one is profitability. During the dot-com era, the ratio of capex spending to free cash flow peaked at nearly four times free cash flow in 2000, meaning companies were spending four times more than they earned. Today that ratio is below one. The biggest AI companies are largely spending what they earn, not what they borrow.
Nvidia posted $215.9 billion in revenue in its last fiscal year with gross margins of 71% and net margins of 53%. By contrast, the dot-com darlings of the late 1990s were burning cash with no clear path to profit. The foundational AI infrastructure companies actually have businesses.
The adoption curve is also different. The internet in 1999 had around 250 million users globally. AI tools have hundreds of millions of active users already and enterprise adoption is accelerating faster than the internet did at a comparable stage.
The Real Risk Nobody Is Talking About
Here’s the honest concern that gets buried in the optimism.
The number one question from investors in early 2026 is whether increased investment in AI is creating a bubble. The technology itself is likely transformative. The question is to what degree and how long before meaningful productivity gains actually show up. That is difficult for anyone to know.
We covered this two weeks ago. OpenAI just missed its own revenue and user targets. Its CFO is worried about paying computing bills. ChatGPT’s market share has fallen from 87% to 68% in 12 months. The most funded AI company in history is already showing cracks.
The dot-com bubble’s most instructive lesson isn’t that the internet was a bad idea. It’s that even transformative technologies can generate catastrophic losses for investors who buy at the wrong price at the wrong time.
Half a trillion dollars in AI capex in 2026. OpenAI missing its targets. The Buffett Indicator at 210%. These things can coexist with AI being genuinely transformative.
The dot-com crash didn’t happen because the internet failed. It happened because prices got so far ahead of reality that any disappointment became catastrophic. The question for AI isn’t whether the technology works. It’s whether the valuations make sense if the revenue takes longer to arrive than expected.
The Bottom Line
The AI boom is real. The technology is transformative. The companies at the centre of it, Nvidia, Microsoft, Google, are profitable and growing fast. This is not 1999, and the bulls have legitimate points.
But the Buffett Indicator is at 210%. Price-to-earnings ratios exceed the dot-com peak. Half a trillion dollars is being spent on infrastructure based on revenue projections that OpenAI, the most important company in the space, is already missing.
The internet changed the world. It also wiped out 80% of the Nasdaq and destroyed $5 trillion in wealth before it did.
AI might do both
