By the end of the 90s, Internet stocks had become a must-have. For many investors, the Nasdaq was poised to replace the S&P 500 and the Dow Jones as the benchmark for US equities. Even if Internet stocks reached sky-high P/E multiples (or even no multiple at all, as many companies were losing money), most investors justified these valuations by projecting very high growth rates for many years to come. These projections turned out to be far too optimistic.

Almost 25 years later, history seems to be repeating itself. The rise of artificial intelligence suggests massive spending on servers, semiconductors and "platforms" such as Meta. However, the phenomenon of market concentration in this field is even more pronounced today than it was at the beginning of the century. Whereas the 5 largest stocks in the index represented 18% of the S&P 500 in the early 2000s, the "top 5" now account for 25% of the index, an all-time high.

Source : DB

Since ChatGPT-3.5 was made public by Open AI in November 2022, the Magnificent 7 have been seen by investors as the sole beneficiaries of the AI revolution, not only because of their ability to attract talent, but also because of their financial power, which enables them to make huge R&D investments in their AI branches. This is an important difference from the dot-com bubble of the 2000s, when it was not just large caps that were riding the ".com" wave; indeed, numerous IPOs of internet stocks had broadened the spectrum of available investment opportunities.

While some AI-related stocks have the potential to deliver a growth rate in line with current market expectations, history tells us that most will not. As explained in a recent article by Research Associates, picking future AI winners is the equivalent of investing in Amazon, Apple and ADP in early 2000. These three stocks were the only ones among the 40 biggest names in the technology sector at the turn of the century to record double-digit performances over the following 23 years.

One of the most marked contrasts is to be found in the valuation and structure of the market. The expected 12-month price/earnings ratio for the Magnificent 7 is 30x, a 50% premium to the rest of the market. These multiples are well below those prevailing in the dot-com era.

In contrast to the dot-com era, large-cap technology stocks are already highly profitable and continue to grow at a strong pace.

The outperformance of the Magnificent 7 since the beginning of the year is supported by earnings revisions. As the chart below shows, the Magnificent 7's earnings expectations for Q4 have been consistently revised upwards, in contrast to the rest of the market.

Another major difference with the years 1999-2000 is capital market activity. The late 1990s saw a frenzy of IPOs, often involving companies with negative cash flows, which ultimately proved unviable.