As readers may have noticed, some investors are anxious that stock market valuations in the US, for example, are becoming overstretched. The vast sums invested in AI are hard to square with likely revenues. Something’s “got to give,” as it were. So, are we in bubble territory or not?
To get a handle on this question, we publish this article from Javier de Berenguer (pictured at the bottom of this article), analyst and fund selector at ratio less meaningful). In such cases, it is more appropriate to use multiples based on sales volume (e.g. price/sales). For more stable companies with mature businesses, the P/E ratio can be a good reference. Similarly, for businesses with significant tangible book value – such as listed financial services or real estate – P/E and price/book (P/B) ratios are reliable benchmarks.
Once this is understood, let us put the American market into numbers – the main focus of valuation concerns – and divide it into three groups: the S&P 500, the S&P 500 Equal-Weighted (EW), and the “Magnificent Seven.”
What the fundamentals tell us
As noted at the beginning, these multiples are useful if we know how to use and, crucially, how to interpret them.
When looking at long-term valuation data (over 30, 15 or 10 years), all three groups – the S&P 500, the S&P 500 EW, and the Magnificent Seven – are trading at P/E ratios above their historical averages.
Although the market-cap-weighted S&P 500 shows the largest deviation from its average, the broader market (not just technology and the Magnificent Seven) also presents valuation levels that can be described as demanding.
If we add fundamentals into the equation, the conclusions shift. Both return on equity (ROE) and net margins for these groups are currently above their historical averages and, in general terms, are expected to continue improving. The market-cap-weighted S&P 500 clearly shows higher returns and margins than the broader market. For example, the S&P 500 has an ROE of around 19 per cent and a margin of roughly 11 per cent, while the equal-weight equivalent shows noticeably lower profitability. This alignment between higher fundamentals and higher valuation multiples helps explain part of the premium assigned by the market.
Therefore, based on these data, the index would be trading at a P/E premium of around 41 per cent over its historical average. It also has ROE and margin figures that are higher and roughly in line with that premium – factors which, taken together, could justify the valuation. However, it is worth remembering that, as mentioned earlier, the reliability of any multiple depends on the sector. So the question remains: has the American market genuinely improved, or has its composition simply changed?
What the composition of the index tells us
Reviewing the evidence so far: we have a US market-cap-weighted index that has both increased its valuation and improved its fundamentals. The broader market, on the other hand, exhibits a balance we might describe as “familiar” when relating P/E ratios to business metrics. The gap in valuation and fundamentals between the two indices has widened considerably in recent years and has become more pronounced with the emergence of artificial intelligence, which has supported the technology sector while the rest of the market has struggled with inflation, geopolitical tensions, supply chain disruptions, the trade war and more.
Over previous decades, the S&P 500 has steadily evolved into an index with a much larger weighting in technology (rising from around 10 per cent in the late 1980s to roughly one-third today) and in communication services. This shift means that the index is now composed of businesses with lower physical capital requirements, higher cash generation and greater scalability – characteristics that naturally justify higher valuation multiples than were typical in earlier periods.
Having examined the peculiarities of the US market, we now move on to comparing the US stock market with the rest of the world.
Compared with the rest of the world – expensive or cheap? And how does this compare with past bubbles?
Contrary to what one might think, there are currently few safe havens in terms of indices or aggregated markets, even when the focus moves outside the US.
When comparing today’s valuations and fundamentals with those seen during previous market bubbles, the figures of the largest companies in today’s index remain far from the extremes observed during episodes such as the dot-com era. The risks today stem more from the concentration of market capitalisation in a small number of companies, and from the fact that passive investment vehicles leave many investors heavily exposed to this narrow cohort.
A comparison with historical bubbles (including the dot-com period) also shows that, while valuations are elevated, they are not yet at the levels associated with late-stage speculative excess. Fundamentally, today’s leaders are generating much stronger earnings and profitability than the companies that dominated earlier bubble periods.
But be careful – there are always risks
Given everything we have reviewed, we can draw both “bad” and “good” news. Starting with the positive: we are not in a bubble, and the evolution of indices currently depends on continued growth in earnings and other metrics (ROE, margins) supporting stock prices. On the negative side, there is no valuation buffer: the market is somewhat expensive, and there is no shame in saying so.
However, several factors could continue to support the markets: a broader group of winners (more companies contributing to overall index returns), efficiency improvements brought by artificial intelligence, or the start of a new earnings growth cycle in industries that have lagged. Macro variables also appear favourable: abundant liquidity, stable economic growth, increased support from monetary policy, and significant public investment commitments (defence, AI, digitalisation, energy and so on).
Another negative point, as noted, is concentration: the lack of a valuation buffer combined with a highly concentrated market pushes us towards increased diversification in portfolios, making it a key tool for managing these risks.
Regarding artificial intelligence, we must pay close attention to how its development unfolds, as this will determine whether we ultimately end up in bubble territory. We are only two years into the race to master this new technology. There is still a long way to go in terms of both the volume of investment required and the number of participants likely to enter. The fact that AI is a transformative technology attracting significant capital is precisely why close monitoring is necessary – these characteristics were also present in markets that eventually ended in bubbles.
Several classic bubble indicators are worth watching: rising participation, rapid inflows of capital, the emergence of new speculative players, and increased use of leverage to finance growth. Some of these elements are already visible in today’s AI investment cycle; others are in their early stages but show signs of accelerating.
To conclude, particular attention should now be paid to financing trends and the growing number of participants. Until recently, hyperscalers (mega-tech) were financing AI development through cash flow, but this is beginning to change: Alphabet and Meta have recently announced bond issues worth several trillion dollars to finance AI. Meanwhile, the concentration of technology among a small group of American companies is decreasing as other players in the semiconductor value chain join the investment wave, including internationally – particularly in China, where major efforts are under way to build a powerful ecosystem around the sector.
The author
Javier de Berenguer
The author is based in Madrid.