AI boom not an internet bubble 2.0

The current strong share price gains in stocks benefiting from AI are based on a more solid foundation than the internet boom at the turn of the millennium.

Thomas Becker

Commerzbank Economic Research

March 22 2024

This is because the most important players are already profitable and much stronger financially than the companies 25 years ago. In addition, the necessary infrastructure is already in place on a large scale. However, the short-term economic implications of the AI megatrend could be overestimated, while the long-term effects are underestimated, meaning that there are likely to be setbacks.

Internet bubble 2.0?

The global stock markets currently seem to be heading in only one direction: Upwards! A key driver for share prices is the hope that the major advances in the development of AI will lead to a significant increase in corporate profits in the medium term. This applies in particular to the seven companies ("Magnificent 7", M7), which are generally expected to benefit the most and whose share prices have risen by around 100% on average since the beginning of 2023. The development of their share prices reminds some people of the time around the turn of the millennium and the "Neuer Markt", when the share prices of companies whose business model was geared towards the use of the internet shot up, only to plummet again after the bubble burst when they failed to meet the high expectations.

This time too, some hopes could be dashed ...

The long-term effects of a new technology are often underestimated, but the short-term effects are often overestimated. The latter is also likely to be the case with AI and can repeatedly lead to temporary share price setbacks.

... but key players much more stable ...

However, we believe that a slump like the one that followed the bursting of the internet bubble or the "Neuer Markt" is almost unlikely. This is because, unlike the Internet companies that were the focus of attention almost 25 years ago, which had early-stage business models and weak capital resources, the AI cycle is being driven by financially strong and globally successful companies. The M7 are market leaders that have already proven their ability to successfully implement new technologies. In view of their considerable financial resources, they should be able to make the necessary massive investments in IT infrastructure and research & development to bring trustworthy AI applications to market maturity.

The companies in the internet bubble of the 2000s were focused on generating sales growth. The profitability of products, projects and customers was only a limited priority. This led to weak balance sheets, negative profit revisions and disappointed market expectations, which together caused share prices to plummet. It is true that leveraging new sales potential also plays an important role for AI profiteers. However, both providers and users of AI applications are also focusing on increasing productivity and profitability.

... and the required infrastructure better developed

Another advantage for AI companies is that they can rely on a developed IT infrastructure. Today, the globally interconnected IT infrastructure in conjunction with the increasing computing power of data centers, the Internet and the availability of digitized content (audio, video, images, text, etc.) forms the foundation on which basic technologies such as AI can spread rapidly around the world. In contrast, the Internet pioneers of the 2000s were only able to realize their visions years later due to a lack of technological foundations at the time of invention.

Valuation less ambitious than 25 years ago

Another difference to the internet bubble is that, despite the rally observed since the beginning of last year, the M7 shares are not valued nearly as highly as comparable stocks at the turn of the millennium. The massive price increase since the beginning of 2023 is put into perspective by the fact that it followed a very weak performance in 2022. In addition, the M7's price/earnings ratio (P/E ratio) of 31 is well supported by expected and realistic annual sales growth of 12% (2023-25). In comparison, the ten most highly capitalized companies, which significantly influence the market indices due to their market capitalization, were valued at a P/E ratio of 47 at the peak of the internet bubble in March 2000. However, the expected annual sales growth of 15% (1999-2001) did not materialize in the end.

For full text see attached PDF-Version.