The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Torsten Sløk, chief economist at Apollo Global Management, warned in July 2025 that the largest companies in the S&P 500 were more richly valued than the market’s biggest companies were during the 1990s information-technology bubble. The comparison has been widely summarized as “the AI bubble is worse than the dot-com bubble,” but Sløk’s claim was narrower: it focused on the valuation of the index’s 10 largest companies, not every AI business or the entire economy.
For individual investors, that distinction matters. High valuations can increase the size of a potential loss, but they do not identify the exact day a market will fall—or prove that a crash is inevitable.
What the economist actually warned
Apollo Academy published Sløk’s note, “AI Bubble Today Is Bigger Than the IT Bubble in the 1990s,” on July 16, 2025. Its central statement was:
“The difference between the IT bubble in the 1990s and the AI bubble today is that the top 10 companies in the S&P 500 today are more overvalued than they were in the 1990s.”
Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.#1 Best Overall
Apollo’s chart compared the 10 largest S&P 500 companies by market capitalization with the rest of the index. It used valuation data available in July 2025 and cited Bloomberg and Apollo’s chief economist.
The companies most associated with the discussion included Nvidia, Microsoft, Apple, Amazon, Meta and Alphabet. However, the analysis was not a ranking of 10 “AI companies.” It was a comparison of the largest companies in the broad S&P 500. Some have substantial AI exposure; others have businesses that extend well beyond artificial intelligence.
“More overvalued” also needs a precise reading. It means investors were paying a higher price relative to earnings under the measure shown in the analysis. It does not mean those companies had no earnings, that their businesses were fraudulent, or that a market collapse had already started.
Why valuation matters to ordinary investors
A price-to-earnings ratio compares a company’s share price with its earnings per share. If a stock trades at 40 times earnings, investors are paying $40 for each dollar of current annual earnings. A high multiple can be justified if earnings grow rapidly and remain durable, but it leaves less room for disappointing results.
For example, suppose a company earns $5 per share:
| Share price | P/E ratio | What happens if earnings stay at $5 and the market pays 25 times earnings? |
|---|---|---|
| $100 | 20 | The valuation would rise to $125 |
| $200 | 40 | The valuation would fall to $125 |
The second example does not require earnings to collapse. The share price falls because investors decide that a lower valuation multiple is appropriate. In real markets, earnings, interest rates, growth expectations and valuation multiples move together, so the result is less tidy than the illustration.
For investors holding a concentrated position in a few technology stocks, a valuation reset can be especially painful. Someone investing through a broad index fund has diversification, but that protection is not complete: the largest S&P 500 companies represent a large share of many market-cap-weighted funds.
The dot-com comparison has limits
The late-1990s technology boom ended with the Nasdaq Composite losing roughly 78% from its March 2000 peak to its October 2002 low. Many internet companies failed, while some surviving businesses eventually became highly valuable. The collapse of share prices did not mean the internet was useless; it meant investors had paid too much for many companies and expected too much, too soon.
That is the useful part of the current comparison. AI can become an important and profitable technology while AI-linked stocks still be overpriced. A bubble can form around a genuine innovation.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →But the current boom is not identical to the dot-com period:
- Many of today’s leading technology companies have significant revenue, cash flow and established customer bases.
- The 1990s comparison concerns selected market valuations, not the entire AI economy.
- AI investment includes profitable semiconductor, cloud, software and advertising businesses, not just unprofitable start-ups.
- Different valuation measures can produce different conclusions. A February 2026 Fidelity review said the S&P 500 and its technology sector were above historical averages but still below late-1990s extremes under the measures it examined.
These differences do not eliminate risk. They show why the headline should be treated as a warning about expectations rather than a precise forecast of a repeat of 2000.
AI revenue is growing, but spending is much broader
S&P Global Market Intelligence projected the generative-AI software market to grow from an estimated $16 billion in 2024 to $85 billion in 2029—about a 40% compound annual growth rate. It also reported that the eight largest vendors held 63% of the generative-AI software market in the second quarter of 2025, while the number of vendors generating more than $10 million in revenue rose from 78 in June 2024 to 138.
Those figures support the idea that demand is real and that the market is expanding. They do not automatically justify every company’s valuation.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteIt is also misleading to compare the $85 billion projected software market directly with the capital spending of individual technology companies. S&P Global’s figure covers packaged software in which generative AI is central to the product. Corporate capital expenditure can cover data centers, processors, networking equipment, electricity infrastructure and other systems used for a much wider range of services.
A company can spend heavily on infrastructure today because it expects future demand. The investment becomes a problem if customers do not generate enough revenue to cover the cost, if computing capacity becomes obsolete quickly, or if competitors add so much capacity that prices fall.
What could make the warning matter
The risk is not simply that AI stops working. A more plausible market problem is that the financial returns arrive more slowly than investors expect.
- Revenue growth disappoints. Businesses may experiment with AI without paying enough for providers to recover infrastructure costs.
- Profit margins shrink. Competition can force down prices for model access, cloud services or AI-enabled products.
- Capital spending overshoots demand. Data centers and specialized chips can be difficult to redeploy if demand forecasts prove too optimistic.
- Interest rates stay high. Higher rates generally reduce the present value investors assign to distant future profits, which can weigh more heavily on high-growth stocks.
- Investors become less willing to pay premium multiples. Even companies that continue growing can see their share prices fall if the market applies a lower valuation.
Later analysis added some nuance. In October 2025, IMF Chief Economist Pierre-Olivier Gourinchas said an AI investment boom could lead to a dot-com-style market bust, but was less likely to cause a systemic financial crisis because the investment was being funded mainly by cash-rich technology companies rather than by the kind of heavy borrowing associated with the 2008 property crisis. The IMF said AI-related investment had increased by less than 0.4% of U.S. GDP since 2022, compared with a 1.2% increase from 1995 to 2000.
Recommended Free Tools
That does not make a stock-market decline harmless. It suggests that a correction could hurt portfolios and business investment without necessarily producing a banking crisis on the scale of 2008.
How personal investors can respond
Sløk’s note was not an investment recommendation, and it did not supply a crash date or a probability. Investors should therefore avoid turning the warning into a demand to sell everything. A practical response is to check whether a portfolio depends on one narrow outcome.
- Review concentration. Add up direct holdings in major technology companies and indirect exposure through S&P 500, Nasdaq or technology-sector funds.
- Check the time horizon. Money needed within the next few years generally should not depend heavily on volatile growth stocks.
- Use a written allocation. Decide how much belongs in stocks, bonds and cash before a large price move makes the decision emotional.
- Rebalance rather than chase. If one sector grows far beyond its target weight, periodic rebalancing can reduce concentration without requiring a prediction about the top of the market.
- Separate the technology from the stock. A company can be a leader in AI and still be a poor purchase at an excessive price.
- Do not use leverage to buy the theme. Borrowed money can turn a normal correction into a forced sale and a permanent loss.
For someone using a diversified retirement fund, the appropriate action may be no action beyond checking that the fund matches the investor’s risk tolerance. For someone whose portfolio is dominated by a handful of AI-linked shares, reducing concentration gradually may be more sensible than trying to predict the exact peak.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the headline does not establish
- It does not prove that the entire AI sector is in a bubble.
- It does not prove that current valuations are higher than the dot-com peak under every measurement.
- It does not predict an imminent crash.
- It does not show that AI technology will fail.
- It does not mean a correction would necessarily become a 2008-style financial crisis.
The strongest conclusion is narrower: the market’s largest companies were carrying unusually high expectations, and investors should recognize that strong technology adoption does not guarantee strong returns from any price.
Best Value
FAQ
Did the economist predict an imminent AI crash?
No. Torsten Sløk warned that the 10 largest S&P 500 companies were more overvalued than their counterparts during the 1990s IT bubble. His July 2025 note did not give a crash date or a quantified probability.
Is the AI bubble really bigger than the dot-com bubble?
That depends on what is being measured. Sløk’s comparison covered the valuation of the S&P 500’s 10 largest companies. It did not prove that the whole AI economy, every AI company or every market measure was more extreme than in 2000.
Could AI still be valuable if AI stocks fall?
Yes. A market correction can reduce inflated valuations, eliminate weak companies and slow infrastructure spending while the underlying technology continues to be useful and commercially important.
Should investors sell all technology stocks?
The warning alone is not a reason for everyone to sell. Investors should review concentration, time horizon, risk tolerance and valuation, then consider diversification and planned rebalancing rather than trying to time a crash.
Free tools Windows power users keep installed
One-click scans. No signup required.
The Bottom Line
Sløk’s warning is best understood as a valuation alert, not a crash forecast. The largest S&P 500 companies were priced for substantial future growth, and disappointment could produce large losses even if AI adoption continues. Personal investors can respond by limiting concentration, avoiding leverage, maintaining a suitable cash and bond reserve, and judging each holding by its price as well as its technology.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
