It's hard to believe today, but in 1999, Warren Buffett's reputation as one of the greatest investors in history was in serious doubt.
For years, he had sat out on the historic tech boom, warning about valuations as the Nasdaq Composite (^IXIC 1.44%) zoomed 85% higher in 1998, then 102% in 1999. That year, Berkshire Hathaway Class A shares fell 23% while the S&P 500 (SNPINDEX: ^GSPC) rose by 18%. Barron's even speculated that Buffett was losing his touch, in a December 1999 article titled, "What's Wrong, Warren?"
But Buffett stuck to his guns. Speaking in Sun Valley, Idaho, before a conference full of venture capitalists, billionaire tech company founders, and dot-com visionaries, he pointed to history to show how overvaluations can destroy returns.
His examples were stark. From 1965 to 1982, the United States economy grew by 400%, yet the Dow Jones Industrial Average (^DJI 1.27%) stayed flat for those 17 years. From 1919 to 1939, when aviation was clearly the future, all 200 airplane stocks averaged out to a 0% gain. And of the 2,000 car companies that sprang up with the rise of automobiles that century, only three survived.
Image source: The Motley Fool.
History was clear: When valuations are excessive, the odds are against investors, no matter how transformative the technology behind the investment is.
Within months, Buffett would be vindicated, as the Nasdaq began a slide that would amount to a 77% crash by 2002. For one chief executive in in the audience, Bill Gates, Buffett's warning would prove especially poignant. Shares of Microsoft took 16 years to recover from the dot-com collapse.
With artificial intelligence (AI) overvaluation fears in the news, Buffett's 1999 warning may be pertinent today. So, are AI stocks overvalued, and therefore dangerous, like internet stocks were in 1999?
A tale of two tech titans
In January 2000, Cisco (CSCO 0.38%) was the world's most valuable company thanks to its reputation as "the backbone of the Internet" through its hardware business. Like semiconductor giant Nvidia, it came to symbolize the tech revolution that had lit a fire under markets.
By January 2000, Cisco was a decade out from its 1990 initial public offering. In those 10 years, its shares returned 30,000%. But there was a red flag in the company's fundamentals: its price-to-earnings (P/E) ratio of 200.
Markets are forward-looking, so a stock's outlook isn't about what the company has done in the past, but what it will do from here on out. Stocks trade at a multiple to earnings, and the P/E ratio offers a snapshot of the deal you're getting should earnings stay constant. Historically, the average P/E ratio for the S&P 500 has averaged around 15 to 20, meaning that if earnings stayed constant, it would take investors 15 to 20 years for those companies to earn enough per share to make back their initial investment.
Of course, earnings don't stay constant over time, so if earnings growth is expected, stocks trade at higher multiples. The danger with Cisco was that, in January 2000, its P/E ratio of 200 priced for extreme, off-the-charts earnings growth in the years ahead, which it couldn't deliver. When investors caught on, shares of the world's largest company fell 87% and took 26 years to recover.
Today, there are parallels to the dot-com bubble. Like then, an explosive technology -- artificial intelligence -- is taking markets by storm, with one obvious poster child for the rally. But Nvidia's P/E ratio is nowhere close to Cisco's circa 2000, at 44. It may be over double the S&P 500's historical average, but when you account for Nvidia's 65% earnings growth last quarter, it's a reasonable premium to pay for the shares.
NASDAQINDEX: ^IXIC
Key Data Points
What about other tech stocks? Microsoft's P/E ratio of 34 also seems reasonable for a company increasing its earnings and revenue by double-digit percentages, with formidable advantages and an entrenched position in the software infrastructure space. Apple, another multitrillion-dollar company, has a P/E ratio of 36. It's latest earnings rose by 86%.
Zooming out, the tech-heavy Nasdaq carries an average P/E ratio of 34. That's barely half of what it reached in early 2000.
So are we in a bubble yet?
While nothing is guaranteed in investing, there's little basis in valuations to argue we're in a bubble, unlike in 1999. Of course, a crash could still be imminent, if something happens to derail the expected earnings growth that has been priced into markets. But while caution is always warranted for investors, we don't appear to be in a bubble in 2025.
