Stock market corrections sometimes come suddenly and other times simmer, like a romantic disappointment that ends in a breakup. Or in a bear market in this case. Results released by Nvidia this week appear to have allayed fears of an AI meltdown, and investors briefly indulged in optimism about the stock market rally. But something has already begun to break in the market’s incombustible mood and the rise with which Nvidia’s accounts were greeted quickly turned into a decline in a context of intense volatility. There is already a suspicion among investors that the euphoria surrounding artificial intelligence may be too much, insufficient to justify the current valuations of these companies. And as the end of the year approaches and it is not at all clear whether the Fed will lower rates in December, profit taking is an attractive option.
The risk of a correction in big tech stocks – and its shockwave on the market as a whole – is no longer a taboo, as demonstrated by the sharp drop suffered by Oracle, of more than 20% in the last month, or the statements of top Wall Street executives acknowledging that perhaps the market has gone too far beyond its expectations with artificial intelligence, the undisputed driver of the US stock market highs. Daniel Pinto, vice president of JP Morgan Chase, noted this week that current valuations of technology companies are pricing in “a level of productivity that will come, but it may not be as quickly as the market is pricing in now.” And Goldman Sachs Chairman John Waldron took direct aim at the elephant walking in the stock markets. “Investors are focused on the dynamics of AI: whether we will get the returns on invested capital that the market expects, and whether that is already a given. That’s the big debate,” he warned this week.
What is being decided in this debate is nothing less than the continuation of a stock market rally that survived this year’s trade war and led small American investors to take on an unprecedented level of risk or the correction that could drag the market as a whole lower, with the added danger of a change in the stock market cycle. For the moment, and despite the negative balance of the week, the optimists clearly win. There are very few pessimists, even if they include among their ranks venerated names such as Michael Burry, the investor who anticipated the mortgage crisis. subprime and who took a bearish position on Nvidia and Palantir and decided to close his fund to the rest of investors to continue alone, without clients demanding short-term profits.
The vast majority of managers and analyst firms maintain firm confidence in investments in artificial intelligence and in the continuity of the stock market bull cycle. “Without taking positions on individual companies, we believe that having sufficient exposure to AI-related stocks is critical to building and preserving long-term wealth. Investors should ensure they are present across the entire AI value chain, including the intelligence and application layers, as well as the enabling layer,” defends Mark Haefele, chief investment officer at UBS Global Wealth Management. In its forecast for 2026, the Swiss firm plans to increase exposure to equities and expects a 15% rise for global stocks over the course of the year, with artificial intelligence and technology as profit drivers. Morgan Stanley is also not giddy and expects the S&P to reach 7,800 points by the end of 2026, which would leave an upside of close to 20% from current levels.
Santander Asset Management, which this week also presented its outlook for next year, has US stocks as its preferred asset for next year. “The expansion of the profit cycle goes beyond technology and extends across multiple sectors and company sizes, generating a balance between solidity, diversification and profitability potential,” defends the manager, who excludes a bubble in artificial intelligence and predicts new highs for 2026.
Citi acknowledges that it prefers to talk about an AI boom rather than a “bubble”. He expects the S&P to end the year around 6,600 points and sets a target of 6,900 points for mid-2026, overweighting technology in the U.S. stock market as well as banks and energy. “Despite potential near-term volatility, we remain optimistic over the medium term, with single-digit upside potential through mid-2026. Strong earnings, a relatively resilient global economy and the prospect of further Fed cuts should support market sentiment and allow the market to continue to grow,” explains Citi.
Optimism is the prevailing sentiment among fund managers, according to the survey conducted by Bank of America among 172 professionals managing half a billion euros. The survey shows that the stock market overweight is the highest since February and that the liquidity position is very low, equal to 3.7% of the portfolio. There is no shadow of pessimism until we get into the details of artificial intelligence. This is where fears emerge, with a growing clarity that calls into question this widespread optimism. Therefore, while managers expect the international stock market to be the best performing asset in 2026, followed by the US stock market, the majority (53%) also assure that AI-related stocks are at bubble levels and that the entire stock market is overvalued, with a majority of 63% of respondents.
For the first time in 20 years, managers believe there is overinvestment in this technology and believe that the bubble burst represents the main market risk. Thus, 45% of managers believe AI failure is the number one risk, up from 33% in the October survey. Nonetheless… the bet on the magnificent seven continues to be the majority position in portfolios (for 54% of managers). Contradiction between what you think and what you invest in or do you simply trust despite everything?
Warnings about the risk of an artificial intelligence bubble are continuous and it is no longer surprising to find them in the financial analyzes of institutions such as the IMF or the ECB. “There is no doubt that the optimism that has accompanied the development of artificial intelligence in recent months is fading,” says Enguerrand Artaz, a fund manager at the French firm La Financière de l’Échiquier.
Citi, which remains bullish on technology, draws a red line. In his analysis, he points out that six of the magnificent seven continue to trade above the 200-day moving average. That is, above the closing price of those 200 sessions. If that ratio dropped from six to four out of seven companies, “it would be very worrying and if it were three of the magnificent seven it would mean that we are convinced that the bull market is over.”
Some managers, decidedly more pessimistic about the trend of technology on the stock market, prefer to entrench themselves in more defensive values and sectors, without giving up on the stock market. At Singular Bank, American shares are underweight and at the Spanish manager Nartex Capital it is clear that “the bubble will burst whatever happens”, even if it is not yet known when. Their conviction is such that foregoing AI-related stocks causes them to miss out on the stock market rally this year, taking losses in their fund.
For those who believe the AI giants’ valuations hold up, LFDE issues a final warning. “If there was a bubble in the markets, it would be more of a bubble of concentration than of valuation,” adds Enguerrand Artaz. That is to say, the level of influence of these large values on the market as a whole is such that it would be difficult to remain on the sidelines in the face of a sharp correction. The manager explains that, with the exception of Palantir, which trades at 130 times 24-month forward earnings, valuations in the technology sector, while elevated, are not as exuberant as they were during the Internet bubble or the Japanese bubble of 1990. Microsoft is now trading at 30 times one-year forward earnings, up from more than 60 times in 1999. But unlike then, the ten largest stocks by capitalization represent 42% of the S&P index. 500, up from 27% in 1999, and the top ten is made up almost entirely of technology stocks.
The correction before the end of the year, which many managers believe to be healthy, would eliminate excesses and ensure a more favorable starting point for 2026. A lower level from which to obtain returns in the new year. The temptation to take profits before the end of 2025 in such a brilliant year for the stock market is strong and the disappointment that the Fed could cause on December 10 if it does not ultimately cut rates – futures give only a 40% chance of such a reduction – could be a trigger for further declines. In the background, the debate persists as to whether the multiples and appreciations accrued by AI-related stocks are irrational or not. The warnings, precise or not, are increasingly abundant and once again refer to the past. It was December 1996 when the then Fed president, Alan Greenspan, spoke of the “irrational exuberance” of technological values, triggered by Internet fever, in a bubble that ended up bursting three years later. Time will tell whether the optimists or the pessimists were right.
