A massive warning signal is flashing for the AI boom
Take a look at this chart. It shows that searches for the term AI bubble on Google have been surging recently.Past performance isn’t a reliable indicator of future results.
A sign that investors are starting to question whether the AI boom has gone too far. In fact, if we look at searches for real estate bubble during the 2000s, we can see that the current level of concern around AI has climbed to the same peak we saw in 2006, right before the US housing market topped out.
Across the headlines, talk of a bubble or irrational optimism in the stock market has become hard to ignore with many drawing comparisons between today's AI-driven rally, the housing boom of the 2000s, and the .com mania of the late 1990s.
Nvidia, the face of the AI revolution, has surged by more than 4,000% in just 5 years, climbing to a $5 trillion market cap, making it the largest listed company in the world. To many investors, that performance feels very familiar. Back in the late 1990s, Cisco saw a similar 5-year, 4,000% rally during the .com boom before its share price collapsed by nearly 80%, erasing most of those gains.
It's comparisons like these that are fueling today's AI bubble talk
And it's not just Nvidia. Other major tech firms deeply involved in AI have also seen huge gains. Tesla has surged more than 2,000% over the past 5 years. Giants like Apple, Google, Meta, and Microsoft have each climbed around 300% during the same period. Even the weakest performer among them, Amazon, has still gained about 155% or more than 20% per year.
Some argue that such massive returns can't last forever, and that markets could be setting up for a correction reminiscent of the dot-com era.
These companies, often referred to as the Magnificent 7, are the ones spending the most to build the foundation of AI itself. Since the release of ChatGPT, their combined capital expenditures have surged from around $160 billion to nearly $400 billion. And by 2026, that figure is projected to approach half a trillion.
Most of this spending is going into data centers, semiconductors, and cloud infrastructure—basically, the plumbing that keeps AI running. And as these firms have grown larger and more dominant, they now make up an ever-bigger chunk of the entire market. As we can see in this chart, the top 10% of companies by market value, most of which are AI related, account for about 75% of the total market capitalization. That is the highest level of concentration on record, even higher than during the late 1990s .com era.
Back then, just a handful of large tech names like Cisco, Intel, and Microsoft dominated the market. And of course, that didn't end too well. When so much of the market depends on just a few companies in the same sector, it can make the market more fragile. And that's what's driving a lot of the concern today.
But price performance and market concentration alone don't necessarily prove that we're in a bubble
As billionaire investor Ray Dalio puts it, a true bubble forms when asset prices move far beyond their underlying reality. One way we can see this is by comparing stock price relative to its earnings or the PE ratio.
As we can see on the chart, the Magnificent 7 stocks have seen their PE ratio more than doubled, rising from around 12 times earnings in 2012 to 30 times today. And compared to the broader market, these companies have steadily grown more expensive over the past decade, widening the gap between them and the rest of the S&P 500, which currently trades at 20 times earnings.
If we break it down by individual stocks, the picture becomes even clearer. Nvidia is currently trading at about 30 times forward earnings, and Tesla is trading at around 190 times. And again, for many investors, they echo the late 1990s where companies like Intel were trading near 40 times earnings and Cisco reached 196 times.
But even with valuations stretched and stock prices sitting near record highs, investors are still piling into the same mega cap tech names. According to Bank of America's Global Fund Manager survey, being long on the Magnificent 7 remains the most crowded trade in the world. More than 40% of global fund managers now saying being long these tech names is the single most popular position in markets today.
There are two possible reasons that could explain this
The first reason is the fact that these companies have never been this profitable. This chart shows the profit margins of the Magnificent 7, which have climbed sharply from around 18% to nearly 27% since the launch of ChatGPT.
When we compare this to the rest of the S&P 500, we can begin to understand why these companies trade at such a premium. While the broader market's profit margins hover around 12%, the Magnificent 7 are generating more than twice that level of profitability.
Profit margins are one of the best leading indicators of future growth. When margins expand, it tells us companies have pricing power, operational efficiency, and the ability to reinvest in what's next. On top of that, even though prices have surged, the earnings of these companies have also grown rapidly.
If we zoom out to include the 1990s and look at the overall technology sector, we can see that since 2020, their earnings have grown by around 165%, far outpacing the 100% growth seen in the 5 years leading up to the 2000 .com crash. For many investors, that kind of performance helps justify the higher valuations.
And from this perspective, one could argue that while the market looks expensive, it's not yet in bubble territory as prices are still supported by improving fundamentals.
Yes, valuations are high and market concentration is at record levels. But for now, profit margins and earnings growth are keeping those prices in check.
In fact, if we look at the PE ratio of the overall technology sector, today's levels still appear moderate by historical standards. At roughly 30 times earnings, current valuations are only about half of what they were during the .com bubble when tech stocks traded at nearly 60 times their profits.
From here, one possible scenario is that if momentum holds and earnings keep meeting expectations, this rally could still have room to run. However, it all hinges on growth expectations. If earnings begin to slow or disappoint, those high valuations could unwind quickly, putting renewed pressure on the market.
One thing that could trigger this is if AI doesn't deliver results fast enough. This chart shows that AI adoption among businesses is still climbing. Today, more and more firms are experimenting with AI tools. But if companies start realizing that these tools aren't improving productivity or cutting costs as much as expected, this adoption could slow down and they might start pulling back on spending—especially on cloud services, chips, and data centers. The very things these tech giants depend on for growth.
Ultimately, time will tell which way the balance tips
But for investors, the key takeaway is this. Keep an eye on earnings. They'll tell whether prices are still grounded in reality or if we're drifting into bubble territory.
At Capital.com, we'll keep you updated on market conditions, corporate earnings, and its risks with videos like this one. Explore more insights in our Educational Hub.
