Goldman Sachs says watch these 5 warnings from the dot-com bubble to know if the AI craze is peaking

• Goldman Sachs says that AI stock valuations resemble some of the signals of the late 1990s bubble.
• Analysts note rising tech investment, falling profits, and widening credit spreads as risks.
• Mega-cap tech firms' AI spending and corporate debt levels may indicate a market peak.

Markets are worried about shades of 1999 in today's tech-investing landscape, and while there's a lot of debate about whether AI is a bubble, there are a few signals from history that show what, specifically, investors should be looking out for.

Strategists at Goldman Sachs said they believe the market's AI frenzy risks mirroring the dot-com bubble burst in the early 2000s.

Stocks don't look like they're in their 1999 moment yet, Dominic Wilson, a senior advisor on the bank's global markets research team, and Vickie Chang, a macro research strategist, wrote in a note to clients on Sunday. But the risks that the AI boom will look a lot like the 2000s era craze appear to be growing, they said.

"We see a growing risk that the imbalances that built up in the 1990s will become more visible as the AI investment boom extends. There have been echoes of the inflection point in the 1990s boom lately," the bank wrote, adding that the AI trade now looked the way tech stocks did in 1997, several years before the bubble burst.

Wilson and Chang flagged several warning signs leading up to the dot-com crash in the early 2000s that investors should be on the lookout for.

1. A peak in investment spending

Investment spending on tech equipment and software rose to "unusually high levels" in the 90s. That peaked in 2000, when non-residential investment in the telecom and tech sector rose to around 15% of US GDP.

Investment spending began to tumble in the months leading up to the dot-com crash, per Goldman's analysis.

"Highly valued asset prices thus had significant consequences for real spending decisions," the strategists said.

Investors have been growing wary of mega-cap tech firms' spending spree on AI this year. Amazon, Meta, Microsoft, Alphabet, and Apple are on track to spend around $349 billion on capex in 2025.

2. Corporate profits started to fall

Corporate profits reached a peak around 1997 before beginning to decline.

"Profitability peaked well before the boom ended," Wilson and Chang wrote. "While reported profit margins were more robust, declining profitability in the macro data in the later years of the boom came alongside accelerating equity prices."

Corporate profits look strong at the moment. The blended net profit margin in the S&P 500 for the third quarter is around 13.1%, above the five-year average of 12.1%, according to FactSet.

3. Corporate debt rose rapidly

Companies grew increasingly indebted in the lead-up to the dot-com crash. Corporate debt as a percentage of profits peaked in 2001, right as the bubble was bursting, Goldman's analysis shows.

"The combination of rising investment and falling profitability pushed the corporate sector financial balance — the difference between savings and investment — into deficit," the strategists said.

Some of mega-cap tech firms' spending on AI has been funded debt. Meta, for instance, raised $30 billion in bonds in late October as it doubled down on its AI spending plans.

Most firms today, though, look like they're financing capex with free cash flow, Goldman added. The percentage of corporate debt relative to profits also looks significantly lower than it did at the peak of the internet bubble.

The Fed cut rates

The Fed was in the midst of its rate-cutting cycle in the late 90s, one factor that helped juice the stock market.

"Lower rates and capital inflows added fuel to the equity market," Goldman wrote.

The Fed cut interest rates by 25 basis points at its October policy meeting. Investors are expecting the central bank to issue another 25 basis-point cut in December, according to the CME FedWatch tool.

Other market pros, like Ray Dalio, have warned that the Fed's easing cycle could help inflate a bubble in markets.

5. Credit spreads widened

Credit spreads widened leading up to the dot-com crash, the bank noted.

Credit spreads — which refer to the yield paid on a bond or credit instrument over a benchmark like Treasurys — widen when investors perceive higher risk and demand to be paid more to compensate.

Credit spreads remain at historically tight levels but have begun to widen in recent weeks. The ICE Bank of America US High Yield Index Option-Adjusted Spread rose to around 3.15% last week, up 39 basis points from its low of 2.76% in late October.

In the 1990s, these warning signs appeared at least two years before the dot-com bubble actually burst, Wilson and Chang said, adding that they believed the AI trade still had room to run.

This story was originally published in November 2025.