The dotcom bubble is now an infamous example of stock market overexuberance. Many would argue that the current AI boom is a classic case of history repeating itself. Is this fair, though – and how similar or different is the AI boom to the dotcom bubble?
One of the most challenging aspects of investing is resisting the urge to pile money into the most popular stocks and funds. There’s a lot of fear of missing out, but as history has shown us, this herd mentality can ultimately lead to financial disaster.
Ever since various big tech stocks, particularly the Magnificent Seven, saw their valuations explode in the wake of ChatGPT’s public launch in November 2022, there has always been a chorus of doubters asking if the AI boom is another dotcom bubble. That’s perhaps understandable.
Subscribe to MoneyWeek
Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE
Get 6 issues free
Sign up to Money Morning
Don’t miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter
Don’t miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter
Today (10 March 2025) marks the 25th anniversary of the dotcom bubble bursting. The Nasdaq Composite hit its peak of 5,000 points on Friday 10th March 2000, up 24% since the start of the year.
As the dotcom bubble burst, though, the index plummeted. By November 2000, it had fallen to below 3,000 points.
“The prevailing belief at the time was that capturing ‘eyeballs’ and subscribers would inevitably lead to riches,” says Greg Eckel, portfolio manager of Canadian General Investments. “However, without sustainable business models, uncertainty prevailed, and many ventures collapsed. While technology advanced, the infrastructure and economic foundations were not ready to support the lofty valuations.”
“The early 2000s serves as a reminder of the difficulties investors face when valuing future innovations,” says Brad Holland, director of investment strategy at Nutmeg.
Cisco (NASDAQ:CSCO) – the largest company in the world in early 2000 – saw its share price fall from slightly over $80 in March that year to less than $10 by October 2002. Even today the stock has not regained the levels it traded at prior to the crash.
Investors bought Cisco because, during the dotcom bubble, the hardware it built appeared to be integral to the internet revolution that was sweeping the world. Any business or government that wanted to keep pace with its competition would, it seemed, need to keep buying Cisco products at scale in perpetuity.
If that sounds familiar, that’s probably because much the same argument is today made by Nvidia (NASDAQ:NVDA) bulls. As the picks-and-shovels stock of choice for the artificial intelligence (AI) boom, which has propelled it to, at times, the top of the world’s most valuable company list, Nvidia seems a direct comparison for Cisco’s status during the dotcom bubble.
So, how do the two bull markets – the dotcom bubble and the present AI boom – compare to one another? And what might that say about whether, when and how the AI bubble might burst?
Are we in an AI bubble?
The defining feature of the late-1990s tech exuberance was the appetite that markets had for unprofitable companies.
This is a challenge that investors always face. It’s worth paying relatively high amounts now for shares in a company that have a high chance of increasing their earnings in future. That’s what venture capital investing is all about.
But how much is too much? And how profitable should a company be before investors start to view it as an established business as opposed to a potential growth play?
When Cisco was the world’s most valuable company, it traded at around 200 times its earnings. Investors bet that the insatiable demand for connectivity would drive ever-greater demand for its products.
In a sense, they were right. Cisco made $10.32 billion of net income in 2024, nearly four times the $2.67 billion it made in 2000. That didn’t make up for the massive differential between its profits and the valuation that investors had given it, though. Once sentiment soured and networking spend dried up, the stock slumped, and hasn’t yet recovered.
For comparison, Nvidia today trades at around 38 times its trailing earnings, and 25 times its expected earnings over the next 12 months. These are a little higher than most value investors would like, but they’re clearly a far cry from the wild valuations investors gave Cisco before the dotcom bubble burst.
“A critical difference in today’s tech dominance is the financial robustness of mega-cap IT companies, which are funding AI advancements with substantial free cash flows, unlike the debt-laden dotcom firms,” says Dom Rizzo, portfolio manager of the T. Rowe Price Global Technology Equity strategy. “This financial strength provides a more stable foundation for growth.”
Incidentally, Nvidia’s trailing price to earnings ratio is down substantially from the 147 it peaked at in April 2023. Its share price, though, is higher now than it was then – showing there’s nothing necessarily wrong with a three-figure P/E multiple, as long as earnings growth delivers.
In other words, Nvidia delivered on the hype, because it has a very viable business model.
Nonetheless, there are hints of exuberance in today’s market. Companies like Palantir (NASDAQ:PLTR) and Tesla (NASDAQ:TSLA) trade at over one hundred times projected earnings. History teaches us that stocks like these could suffer losses, should sentiment turn against the AI boom.
So, yes, we are in an AI-driven bubble, but every bull market is a bubble of sorts. The more important question is how and when this period of stock market exuberance might come to an end – or, how the AI bubble might burst.
(When) will the AI bubble burst?
The factors that led to the bursting of the dotcom bubble were complex. It’s not clear that a similar set of circumstances is poised to burst the AI bubble any time soon.
One particularly noteworthy distinction between the two eras is that, in most developed economies, interest rates are currently falling, albeit gradually. That’s a tailwind for growth stocks, and the opposite to what happened in the runup to the dotcom bubble bursting.
Valuations are not particularly high by comparison, either. The S&P 500’s average P/E ratio is currently well below 30, as opposed to during the dotcom bubble when it shot past 30 and even into the high 40s.
“Gaining some perspective is important when looking back at these events, which have shaped global markets and linger in the psyche of investors,” says Holland.
“The hyperconnectivity of today is unrecognisable compared to March 2000 when 2G networks were the norm and new technologies were far less engrained than they are now.
“Looking back, the market excitement and ‘get in early’ mentality, which led to the creation of the bubble, was heavily influenced by hype around companies that had yet to deliver significant shareholder value and impact.”
That said, there’s no reason why any present AI bubble should need to burst in the same way as the dotcom bubble did. There are various other ways that the AI bull market could end.
The appearance of DeepSeek, a Chinese generative AI start-up that appears to be able to outperform many Western AI models for a fraction of the capital and compute costs, certainly shook the market at the start of the year.
Should Donald Trump succeed in plunging the world into a tariff-driven trade war, then the subsequent inflationary shock could prompt a stock market reversal.
Or, as tends to be the case with stock market crashes, the catalyst could be something that catches everyone completely off-guard.
The cautionary tale of the stock market is to keep a close eye on the fundamentals of the stocks you invest in. Solid businesses like Amazon (NASDAQ:AMZN) and Microsoft (NASDAQ:MSFT) weathered the storm that followed the dotcom bubble and are now among the leaders of the AI boom.
It just pays to know, as much as possible, what the risks involved in buying stocks, particularly growth stocks, involve.
