David Einhorn recently warned about the valuation levels reached by the markets, particularly in the artificial intelligence sector. During a panel discussion held last Thursday at the New York Stock Exchange, he warned of the extreme level of spending on AI-related infrastructure by giants like Apple, Meta, Nvidia, Microsoft, and Google, which, according to him, increases the "reasonable probability that a considerable amount of capital will be destroyed during this cycle." Also during the same panel discussion, the founder of Greenlight Capital was able to demonstrate a somewhat gloomy outlook, stating that he "rather believes we are entering a recession, if not already there. Jobs are not growing, the workweek is decreasing, and productivity is not brilliant." While US growth remains solid (+3.3% revised to +3.8% in Q2 annualized), signs of a slowdown are accumulating, particularly in the labor market (unemployment at 4.3% and decline in job creation), and inflationary pressures persist (the Core PCE index, published on September 26, 2025, remained stable at 2.9% year-on-year).
Still, US equities remain supported by the Fed's policy, which is gently ending its monetary pause, the specific premium linked to artificial intelligence, which is becoming widespread, and the weak dollar, which is favoring exports and multinational profits. The euphoria in US equity markets persists, and staying completely away from them seems unconstructive, even value-destructive. In times of excessive valuation, it is indeed tempting to listen to the apprentice Cassandras, always quick to emerge to announce the imminence of a reversal. As is often the case, their warnings only take on their full meaning once the crisis has passed... Certainly, the Nasdaq 100 has reached a capitalization equivalent to 105% of the American GDP and has grown by 3.3 times in five years. Also, if the S&P 500 and especially the Nasdaq 100 experienced a technical respite last week, this does not seem to call into question the underlying upward trend. It is true that this dynamic does not fail to recall the famous warning of Alan Greenspan, then chairman of the Fed, who posed this now famous question in 1996, in a speech at the American Enterprise Institute: "How do we know that irrational exuberance has not caused asset values to rise beyond what is justified?" For his part, Robert Shiller, Yale professor and Nobel Prize winner in economics, published his famous book, "Irrational Exuberance," in March 2000, a few days before the Nasdaq peak, and wrote: "Financial bubbles are not just economic phenomena; they are deeply rooted in human psychology." However, it would be simplistic to believe that Alan Greenspan and Robert Shiller had "predicted" the Internet bubble. Greenspan had mentioned a possible "irrational exuberance" as early as 1996, but he had pursued an accommodative monetary policy, contributing to the expansion of valuations. His warning, although famous, lacked consistency with his decisions, which earned him much criticism. Shiller, for his part, although he rigorously analyzed the market's excesses, had neither proposed a scenario nor anticipated the timing of the burst. His status as a visionary rests largely on the coincidence of the timing, rather than on an explicit forecast! In short, both signaled imbalances, but history has amplified their role, sometimes at the cost of a selective reinterpretation of the facts.
In a context of high valuations and sometimes contradictory signals, it is essential to remain invested with a long-term vision, by being discriminating in sector choices and rigorous in stock picking. Let's simply recall the conclusions of the Hartford Funds study updated in January 2025 (data from Ned Davis Research and Morningstar): missing the 30 best days of the S&P 500 between 1995 and 2024 would have reduced an investor's returns by 83%. Patience and resilience often pay off more than market timing. So, let's drink a large glass of cold water and remember that the best strategy remains patient, selective, and diversified investing.