TradingKey - Against the backdrop of the AI wave continuing to propel U.S. stocks higher, market discussions regarding whether valuations have entered unsustainable territory are intensifying.
Since late March, the S&P 500 has risen a cumulative 13%, including a 15% surge in the two months prior to last Friday's pullback—a gain that ranks in the top 1% of historical returns since 1980.
This rally has been primarily driven by AI-related stocks like Micron Technology and momentum trading strategies, sparking widespread questioning of whether the advance has been too rapid.
Goldman Sachs ( GS) noted in its latest research report that while some signals of overheating are indeed surfacing, multiple key indicators show the current market remains significantly distanced from the extreme levels seen during historical bubble periods.
The report stated: "Speculative mania itself is not a precise timing tool, but it is indeed one of the typical characteristics of the peaking stages of past high-valuation, high-concentration bull markets."
Meanwhile, Citigroup ( C) significantly raised its year-end 2026 S&P 500 price target to 8,100, based on an "unprecedented" AI capital expenditure supercycle.
Goldman Sachs strategist Ben Snider and his team developed a monitoring framework in a report that covers four dimensions—price performance, trading activity, investor sentiment, and corporate expectations—comprising a total of nine indicators.
Research shows that the median ranking of these indicators currently sits at the 86th percentile historically since 1995; in contrast, this metric reached the 100th percentile during the dot-com bubble and the 95th percentile during the 2021 market peak.
The report specifically notes that market breadth has narrowed significantly, meaning that the majority of gains are driven by a relatively small number of stocks. However, Goldman Sachs emphasized that current market concentration remains lower than it was during the tech bubble of the late 1990s. Unlike previous speculative rallies, the recent rise in U.S. stocks has been largely driven by improvements in earnings expectations.
Since the beginning of this year, consensus earnings per share (EPS) expectations for the S&P 500 have risen by 16%, outstripping the index's 8% price gain. Goldman Sachs expects profit growth to remain robust, forecasting that S&P 500 EPS will reach $340 in 2026, a 24% increase from 2025.
Regarding trading activity, Goldman Sachs' speculative trading indicator has risen in recent months but remains below the levels seen during the dot-com bubble and the 2021 market surge.
The indicator primarily tracks trading in loss-making companies, penny stocks, and stocks with high valuation multiples. Among these, trading activity for high-valuation stocks with enterprise value-to-sales (EV/Sales) ratios exceeding 10x has approached its highest level in decades, second only to the dot-com bubble of 2000.
Meanwhile, the median short interest for S&P 500 components stands at 3.2% of their market capitalization, the highest level since the 2008 financial crisis and significantly higher than the levels seen during the 2000 and 2021 market peaks. Goldman Sachs believes this suggests that investors' actual stance is more cautious than many sentiment indicators would imply.
Citi’s perspective focuses more on structural changes in earnings momentum. Entering 2026, the bank's baseline forecast for S&P 500 earnings per share (EPS) was $320, but actual first-quarter performance exceeded market consensus by approximately 13.4%. Historically, this level of outperformance has only been seen in the early stages of post-recession recovery—yet there is no recessionary context at present. Consequently, Citi raised its full-year 2026 earnings forecast to $350 and provided a preliminary 2027 outlook of $400.
Citi explicitly refuses to define the current environment as a "traditional cycle," arguing instead that a one-off capital expenditure (CAPEX) super-cycle is currently in a "mid-cycle phase." While earnings growth momentum has not yet peaked, the stage of most rapid acceleration may have passed. Going forward, P/E ratios will come under pressure, and index gains will increasingly depend on earnings growth itself rather than valuation expansion.
Although Citi raised its price target to 8,100 points, the report is fraught with warnings about a "valuation cliff." Citi clearly stated that the future driver for the index will be earnings growth, not valuation expansion.
In fact, the 8,100-point price target implies a lower trailing P/E ratio than previously. The market has fully identified the "selling shovels" trade logic and has likely priced in AI-related growth through 2027. However, the fundamental transmission from 2028 to 2030—the transition from AI providers to broader AI users and its translation into actual productivity—remains opaque.
The current U.S. stock market is in a delicate environment. On one hand, AI-driven earnings upward revisions and CapEx expansion provide fundamental support for valuations; on the other hand, sticky inflation, expectations of a policy pivot, and marginal tightening of liquidity constitute potential headwinds.
Goldman Sachs believes that while all historical conditions signaling a market peak have not yet fully converged, certain warning signals are indeed intensifying: IPO activity is gradually recovering, corporate profit margins are facing upward cost pressures, and financial markets have begun pricing in the possibility of a Federal Reserve policy pivot.
Citigroup stated bluntly that this is a market with a very low margin for error—the convergence of growth stocks' forward P/E ratios toward their 10-year averages does not require much compression, but because of their extreme weight in index earnings, the impact of such valuation adjustments on the S&P 500 will be magnified.