The AI rally is fueled by high expectations, but the market is becoming sensitive to reality
AI trading has so far been based primarily on the market buying the future, not current results. In this environment, the market often evaluates not only current profits but also the pace of growth, margin expansion, market share, and a company’s ability to stay ahead of the competition. The problem is that for many AI stocks, these expectations are already largely priced in. If a weaker outlook, slower monetization, or even just a slight slowdown in growth then emerges, the market’s reaction tends to be disproportionately harsh. That is why the question today is no longer just whether AI is an important technology, but primarily whether its financial benefits will materialize in a timely manner and to a sufficient extent. Citigroup’s bear market risk checklist tracks 18 warning signs, including valuations, earnings estimates, capital inflows, capital expenditures, sentiment, and stock issuance, which is important because several of these indicators are already beginning to deteriorate simultaneously. The index itself reached 10 out of 18 points, while in the U.S., the indicator was even higher, at 11.5 out of 18, suggesting that the market is no longer in a comfortable phase. With such a score, we are in an environment where vulnerability to negative surprises is increasing, even though the economy has not yet formally entered a recession. This is precisely why the AI narrative is becoming a fragile topic. The higher the valuations, the less forgiveness there is for weaker results. [1]
The biggest weakness may not be the demand for AI, but the cost of capital
The stock market today is threatened not only by waning optimism surrounding AI but also by the cost of capital itself. The interest rate environment is particularly important for growth stocks, as their value is largely based on future cash flows, which are discounted using interest rates. When rates are higher, or the market believes high rates will persist longer, the valuation of future earnings declines, and growth stocks lose support. This is one of the main reasons why the technology sector in particular often comes under pressure when bond yields rise. Investors begin to demand a higher return for waiting for future earnings, thereby changing the entire valuation equation. Bond yields and expectations surrounding the Fed have therefore begun to put very direct pressure on tech stocks. Market bets on future interest rate movements have shifted sharply, and following strong economic data, the likelihood of further monetary tightening has increased. At the same time, the market is also watching the leadership change at the Fed itself, with Kevin Warsh coming to the forefront, which increases uncertainty surrounding the central bank’s future approach. This is crucial for AI stocks, as higher rates affect not only valuations but also investors’ willingness to hold expensive growth stocks that do not generate immediate returns. When combined with a weaker appetite for risk, capital begins to shift toward more defensive market segments. [2]
Capital expenditures are skyrocketing, but returns remain unconvincing
One of the strongest arguments against the boundless AI euphoria is the scale of capital expenditures. Major U.S. tech companies are spending sums on AI infrastructure, data centers, chips, network capacity, and energy security that were unimaginable just a few years ago. Estimates suggest that AI-related investment spending could reach approximately $800 billion by 2026. Morgan Stanley also estimates that global investment in data centers could reach as much as $3 trillion between 2025 and 2028. This is a scale capable of transforming the entire industry as well as broader financial market conditions. At the same time, there is growing discussion about whether this capital can be recouped quickly enough in the form of higher revenues, margins, and free cash flow. Even more importantly, investors are no longer looking solely at the size of the investments, but also at their efficiency. If hyperscalers and major players spend enormous sums, yet revenue from AI services grows more slowly than expected, the market will begin to doubt the return on investment. Both CNBC and Goldman Sachs are working with lower, though still very high, estimates of around $527 to $725 billion for hyperscalers’ capital expenditures in 2026, which only confirms that the market is in an extreme investment phase. The result is simple: the AI boom remains strong, but the room for disappointment has shrunk significantly. [3]
AI can be both a growth engine and a new inflationary problem
There is also a less obvious but very important layer to this story: The AI boom may not only boost productivity but could also create new inflationary pressures. The proliferation of data centers, rising prices for memory chips, higher demand for electricity, and more sophisticated digital infrastructure are increasing the cost base at a time when inflation is not yet fully under control. This means that, from a macroeconomic perspective, AI cannot be viewed solely as a net disinflationary factor. In fact, in the short term, it may have the exact opposite effect, namely, putting upward pressure on input prices and financing costs. This is an uncomfortable situation for the Fed, as a weaker decline in inflation gives it less room to cut rates. Furthermore, oil prices and geopolitical tensions in the Middle East are also entering the macroeconomic picture, increasing cost pressures and reducing the chances of a rapid easing of monetary policy. Prices of goods in the core component of the PCE index rose by an annualized 5.5% from November to March, with software alone contributing 2.8 percentage points to this increase, and USB flash drives becoming 70% more expensive due to a shortage of memory chips. This is a very strong signal that the AI boom is not just a technological phenomenon, but also a price driver in the real economy. Add to this a more hawkish Fed, higher yields, and persistent geopolitical uncertainty, and the AI rally could shift from being the market’s main driver to its most vulnerable component. Ultimately, what matters is not just how much companies invest, but whether these investments translate into revenue and profit growth within a timeframe that is still acceptable to investors. If monetization turns out to be slower than anticipated, the market will begin to reassess not only individual stocks but also the entire sector narrative. That is precisely the moment when a growth theme becomes a market risk. [4]
[1,2,3,4] Forward-looking statements are based on assumptions and current expectations, which may be inaccurate, or on the current economic environment, which is subject to change. Such statements do not guarantee future results. They involve risks and other uncertainties that are difficult to predict. Actual results may differ materially from those expressed or implied in any forward-looking statements.
Sources:
https://www.reuters.com/data/ai-boom-sparks-rally-frenzy-fear-2026-06-10/
