Nasdaq Composite earnings multiples are way below what was seen in 2000.
Technology stocks are not in an artificial-intelligence-induced bubble and attractive investment opportunities in the sector remain, according to Goldman Sachs.
In a strategy paper published Thursday, a team of Goldman strategists led by Peter Oppenheimer noted that it was 25 years ago this week that the dot-com bubble burst and that the anniversary was encouraging investors to note comparisons with current signs of a selloff in big tech.
"Given the correction so far in technology stocks in 2025, particularly in the U.S., we investigate the parallels and differences between the two periods and examine what lessons can be learned," the Goldman team wrote.
Their main finding was that the valuations of dominant technology companies today are much less extreme than those of the dot-com bubble and that the fundamentals of the technology sector today are also much stronger.
Questions about whether today's big tech stocks may founder in similar fashion to the end of the dot-com frenzy have become more frequent as the current bull run has progressed.
Goldman observed that exuberance around the commercialization of the internet saw the Nasdaq Composite COMP rise fivefold between 1995 and 2000, reaching a price-to-earnings ratio of 200. But within a month of its peak in March 2000, the Nasdaq had lost 30% and continued to trundle lower for more than two years, shedding 80% before bottoming.
A sea change in technology has also driven the latest upsurge for the Nasdaq. Since the craze for AI-related stocks was sparked by the November 2022 launch of ChatGPT, the Nasdaq has jumped about 58%.
Goldman noted that several of the factors that drove the cycle in the late 1990s resonate with enthusiasm for AI and its related technologies today. "A sea change in technology seems to be at a critical point of commercialization, bringing the potential for higher future growth," the team wrote.
The difficulty for investors now, as it was as the internet bubble grew, is how to value the economic benefits that will accrue and identify who will be the biggest winners and losers. "Ultimately, bubbles develop as the aggregate value of companies that may be involved in the innovation exceed the future potential cash flows that it is likely to generate," Goldman said.
However, the big difference between 2000 and 2025 is that the rising valuations of most tech stocks are justified by fundamentals - particularly, as the chart below shows, surging profits. The Nasdaq Composite's forward P/E ratio is 25, a quarter of its peak in 2000.
Source: Goldman Sachs
This has left the technology sector's price/earnings to growth ratio, or PEG ratio, which shows a stock's valuation multiple relative to its expected profit growth, at about 1.4, pretty much in line with the rest of the market.
Another difference between now and 2000 is that technology profits have been increasingly concentrated in a small number of dominant companies, meaning the safety of their earnings can attract a higher multiple.
However, there is a risk that new AI-related technologies might disrupt some of these incumbents, Goldman warned.
And the team acknowledged that there is evidence of bubble behavior in some pockets of the market, where speculative nonprofitable growth companies have seen dramatic share-price surges not backed by fundamentals.
Meanwhile, a main risk to the big tech incumbents currently is that their recent increased spending on capital expenditures - notably AI-linked projects - will not produce the return investors expect.
Source: Goldman Sachs
Despite those caveats, Goldman implied that many of the leading U.S. big tech stocks are a better value now given that their P/E multiples have drifted back as they have underperformed the broader market this year.
"The Magnificent 7 now collectively trade at a P/E of 27x, the lowest since early 2023. This de-rating has occurred despite consensus expectations that the group will collectively continue to grow earnings per share at a faster rate than the S&P 493," the team wrote.
Comments