Slok’s report featured a bar chart showing the median P/E multiple for the 10 most valuable U.S. companies at five-year intervals from 1990 to 2020, and as of late February 2024, as well as displaying the overall numbers for the S&P 500, and the index excluding the top ten. His telling headline: “The current AI bubble is bigger than the 1990s tech bubble.”

The source of that assessment? Slok’s graph shows that the top 10 P/E multiple is now hovering at around 40, far above the readings in the 23 to 26 range between 1995 as the dotcom frenzy gathered momentum and the craze’s near-apex in 2000. His conclusion: “The top 10 companies in the S&P 500 today are more overvalued than during the tech bubble.”

Slok’s analysis pointed me to examine the list of stocks that commanded those huge valuations in 2000, those still in the top 10 club over two decades later, and how the individual names, and the batch as a portfolio, have fared through today. I also decided to use a different metric from Slok’s by replacing the median multiples with one based on the combined P/E for the whole group, in other words, their total market cap divided by all-in profits—as if they were one giant enterprise. That way, I could determine how investors would have fared holding a cap-weighted assemblage for the entire big 10, and what that might tell us about the odds that today’s buyers will greatly profit from holding Nvidia and the other trillion-plus tech glamour names, and hence going all in on the AI theme Wall Street’s selling and that so far, Main Street and fund managers are buying in droves.

How stocks that inflated during the ’90s tech bubble have fared since

Most market-followers could probably name the first half-dozen of the top 10 blindfolded, and likely in order. They’re all tech titans and high-fliers: Microsoft ($3.04 trillion), Apple ($2.80 trillion), Nvidia ($1.98 trillion), Amazon ($1.82 trillion), Alphabet ($1.75 trillion), and Meta ($1.23 trillion). The remaining four are Berkshire Hathaway ($895 billion), Eli Lilly ($740 billion), Tesla ($635 billion), and Broadcom ($566 billion). All told, their market cap stands at $15.1 trillion, and over their past four fiscal quarters, they’ve booked $468 billion in GAAP net profits. Hence, their combined P/E by our measure is 32.3. That’s huge, though not as big as the 40 median cited by Slok. The reason for the difference is that the big multiples are concentrated among the top six at between 37 (Microsoft) and 141 (Lilly), while Meta, Alphabet, and Apple are all moderately lower and a single outlier occupies the deep “value” range, namely Berkshire.

In fact, the group reading of 32.3 is deceiving, because Berkshire, as the only old-economy stalwart on the list, sells at just 12 times its profits for the past four quarters. Excluding Warren Buffett’s colossus, the others in the top 10 combined are sporting a far more formidable P/E of 36.

The question that Slok prompted: At valuations rivaling those in the tech bubble, should investors who consider buying some or all of these names proceed with extreme caution? Also concerning is that as these mostly superexpensive stocks get richer and richer, folks who favor index funds tracking or partially tracking the S&P 500 keep plowing more and more of their savings where they get the fewest dollars in profits per dollars invested, and shunning what’s cheap and unloved.

AI stock market predictions

The fate of the top 10 from 2000 suggests that few of today’s winners will prove hugely successful over the next quarter-century, and several may crash from the heights.

Leading the Billboard top 10 in December 2000 was “Independent Women” by Destiny’s Child, followed by the likes of “Kryptonite” from 3 Doors Down, and the Pink classic “Most Girls.” On the market-cap charts, the champ was General Electric at $472 billion. The other nine, in order of valuation: Exxon Mobil, Pfizer, Cisco, Citigroup, Walmart, Microsoft, AIG, Merck, and Intel. Twenty-three years later, the only name still in the elite grouping is Microsoft. The class encompassed just three tech stars versus eight today (including Tesla), and was far more diversified, covering retailing, energy, financial services, and everything from aircraft engines to appliances via GE.

The group’s combined P/E was 33.1, a hair above today’s number of 32.3, but well below the mark of 36, excluding that wild card Berkshire.

Individually, the stars of 2000 performed poorly for the most part over the following two-decades–plus. The worst were Great Financial Crisis victims AIG and Citi. Their stocks are down 94% and 80% from the close on the final trading day of 2000, measured by total shareholder return. In that span, GE delivered a TSR of 0, Cisco managed a positive 29%, Pfizer +42%, and Intel +152%. Merck (+240%) and Exxon Mobil (+385%) did much better, but still lagged the S&P 500’s TSR from the start of 2001 through late February 2024 of 500%.

Put simply, eight of the 2000 top 10 trailed the big-cap benchmark, most by a big, big distance. By contrast, the gang boasted two gigantic winners. Microsoft stock jumped 29-fold, and Walmart multiplied its shareholders’ stakes 11 times. In retrospect, the array acted something like a biotech company’s portfolio of drugs in development, or a movie studio’s roster of action thrillers. In other words, a mixture of mainly turkeys and a couple of blockbusters that hopefully will pay for all the clinkers and then some.

How much did the winners contribute to bail out the eight laggards? Overall, weighted by market cap, the 10 from 2020 through February of 2024 gained just over 400% in TSR, thanks overwhelmingly to the powerhouse performances from Microsoft and Walmart. That’s an annual return of 7%. The distressing part: The S&P’s 500% increase equates to a yearly return of 8% that beat the 2000 vintage, big 10 result by a substantial one percentage point annually, culminating in a 25% beat for the index over the full period.

What can we learn from comparing the top 10 from the tech bubble days to their 2024 counterparts? First, today’s members are really, really expensive versus history, as Slok points out. Seven are selling at P/Es of between 31 (Meta) and 141 (Lilly), and Apple at 28 times earnings that are at almost twice pre-COVID levels, and barely growing, also looks sumptuously valued. Only Berkshire appears, by traditional standards, to be selling at an attractive P/E. Second, if you’re tilting toward the eight tech superstars that dominate the class of 2024, you’re making a highly concentrated wager on the future of AI. More than anything else, it’s the buzz over how the world-altering new technology will raise the U.S. tech giants’ profitability to a whole new level that’s, since the start of October, driven the P/E for the contingent, sans Berkshire, from 29 to 36. These stocks are regarded much more as sharing a common narrative than the highly varied bunch in 2000. Which makes buying them in concert even riskier.

It’s also important to consider how the top 10 crew is increasingly the tail wagging the market. It’s their soaring valuations that have lifted the S&P 500 multiple from 23.3 at close of last year’s Q3 to almost 27 by late February. By the way, a 27 P/E slightly exceeds the gauge of 26.4 at the end of 2000. Two years later, the index had slumped by 33%.

At these prices, if AI doesn’t pan out, today’s top 10 could do even worse than the group from 2000. Will another Microsoft or two rescue the losers? The best candidate, we keep hearing, is Nvidia. And if you want “scary,” run the numbers on what the AI phenom must do just to give decent returns on its own, let alone prove savior for a basket of flops. Wall Street wants you to think that the fun for these marquee revelers is just beginning. Instead, the failure of AI to deliver as big or fast as Wall Street deems may prove the kryptonite that busts the blast: 3 Doors Down on the turntable chanting their 2000 hit would furnish a fitting coda as the bash winds down.