Does a financial bubble really exist in the AI sector? Questions and answers

Italian public debt stands at around EUR 3,000 billion, less than Nvidia’s market capitalization. What are the signs that suggest an artificial intelligence bubble may be about to burst?

Prepared by XTB

Names such as Nvidia and Apple have topped USD 4,000 billion in market capitalization, Alphabet (Google) and Microsoft are above USD 3,000 billion, while Amazon is above USD 2,000 billion; only Meta stands at USD 1,500 billion. These are market capitalizations that are completely out of scale, especially when compared with the rest of the market and, above all, with long-established rivals such as Amd (USD 348 billion) and Intel (USD 176 billion).

These market capitalizations of American tech stocks can also be easily compared with European names such as the much-discussed ASML, the global supplier of the machinery used to build the chips and hardware these companies produce and use, a company with a market capitalization of “only” USD 400 billion (around EUR 350 billion). Or we could look at public debt, such as Italy’s, which comes to around EUR 3,000 billion, less than the market capitalization of Nvidia, Apple and Google individually. The comparison becomes even more striking if we also consider stock indexes and entire markets, such as the Italian equity market, which is worth just over USD 1,000 billion with 143 listed companies, or the German market at just over USD 2,800 billion with 375 listed companies. The gap that big tech AI names have over other markets should absolutely be seen as a key factor when thinking about a bubble in the AI sector.

The Buffett Indicator relates the total value of the US equity market (Wilshire 5000 index) to US GDP, a relationship that shows how large the equity market, and therefore the financial sector, is compared with US GDP, and thus with the economy as a whole. Historically, this measure has been used to determine whether markets were overpriced or underpriced. A reading above the 150% area could already indicate an overvalued market pricing in relatively high risks. This 150% level was reached during the 2000 dot-com bubble, corresponding to the second standard deviation. The standard deviation for this indicator is important because it tells us where we actually stand in terms of overpricing or underpricing. We are now at 2.4 times the standard deviation, therefore above the levels of the dot-com bubble with a ratio well beyond 150%; we are at around 220%. Looking at this indicator alone, the situation has been critical for some time. Be careful: this indicator tells us about the current state of affairs but is not predictive when it comes to technical market moves.

The undisputed queen of the AI sector is surely Nvidia, which over the past few years has seen absolutely astonishing gains in terms of market capitalization: it was the first company in the world to reach a market capitalization of USD 5,000 billion. Its revenue continues to grow; in this case, we are talking about growth of more than 100% from 2023 to 2024, from about USD 60 billion to USD 130 billion, while projections for this year point to growth toward USD 200 billion, meaning growth of between 50% and 60%. Very strong growth, but with a slight physiological slowdown compared with previous years, a deceleration that certainly does not point to underlying structural weakness. What could affect such a slowdown may be the arrival of Nvidia competitors, who could erode part of the market share Nvidia holds in hardware. This assumption comes from the performance of tech stocks since the start of the year.

Taking Nvidia as the reference “benchmark,” companies such as AMD, Intel and Alphabet (Google) have performed very well, with returns far stronger than Nvidia’s roughly 30%: Intel +82%, AMD +78%, Alphabet (Google) +69%. AMD and Intel are starting to make themselves felt, precisely Nvidia’s historic rivals, while Alphabet (Google) has recently joined in, trying to compete at the hardware level. The other megacaps, meanwhile, remain below Nvdia in year-to-date performance, with Microsoft at just under +17%, Apple at +15%, Meta at +6% and Amazon at +4%.

We are nevertheless seeing a strong tendency among these companies to develop data centers, AI hardware and software. The sector is still expanding rapidly from a fundamental standpoint, but the gap between market prices and the underlying numbers for the various competing companies is clearly skewed. Even indicators such as the Buffett Indicator point to a situation that must be approached with extreme caution from a fundamental perspective: markets have never recorded such high levels of overpricing. From a technical standpoint, markets still do not show signs of weakness sufficient to identify the bursting of a bubble, and on that front major indexes such as the Nasdaq and S&P500 may provide us with signals in the future, especially because of the enormous weight these stocks have within the indexes in terms of market capitalization.