AI valuations now higher than 1990s dot-com bubble, Apollo economist warns

Apollo Global Management chief economist Torsten Slok has warned that the current AI bubble is more dangerous than the conditions leading up to the dot-com crash of the late 1990s. According to Slok’s analysis, the top 10 companies in the S&P 500 today show higher price-to-earnings ratios than they did during the infamous tech bubble, suggesting severe overvaluation in AI-heavy stocks.

What you should know: The market’s AI frenzy has created valuation levels that exceed even the notorious dot-com bubble era.

• “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,” Slok wrote in a widely circulated note.
• Price-to-earnings ratios for the top ten S&P 500 companies have climbed steadily over the past five years, reaching levels well above those seen throughout the 1990s.
• AI chipmaker Nvidia leads the S&P 500 by index weighting, followed by Microsoft, Apple, Amazon, Meta, and Alphabet—all companies that have made multibillion-dollar AI investments.

The big picture: Tech giants are pouring astronomical amounts into AI infrastructure while earnings struggle to keep pace with spending.

• Despite explosive popularity for tools like ChatGPT and Google’s Gemini, revenue remains modest compared to the tens of billions being spent on data center expansions.
• Meta alone will spend over $60 billion on capital expenditures this year, while S&P Global projects the entire generative AI market will reach only $85 billion in aggregate revenue by 2029.

Why this matters: The disconnect between AI spending and actual returns has created conditions that multiple experts compare to historic market crashes.

• AI critic Ed Zitron has drawn parallels to the subprime mortgage crisis of 2007, which collapsed the US housing market.
• Even in 2023, when ChatGPT was only months old, analysts warned investors were overinvesting in unproven technology.

Market vulnerability exposed: Earlier this year, Chinese AI company DeepSeek demonstrated how a fraction of typical computing costs could produce competitive AI results, triggering over $1 trillion in market selloffs.

• DeepSeek’s chatbot matched performance with large language models from OpenAI, Meta, and Google while requiring significantly less computational power to train.
• The incident revealed Silicon Valley’s vulnerability to more efficient AI approaches that could undermine massive infrastructure investments.

What’s at stake: The sustainability of current AI valuations depends on whether companies can eventually justify their enormous bets on the technology.

• Tech companies face mounting pressure to demonstrate that their AI investments will generate returns proportional to spending levels.
• The outcome will determine whether the AI boom continues or becomes “the sequel of one of the biggest stock market crashes in history,” as the analysis suggests.