"Are we being too optimistic?" This is the question Goldman Sachs strategists have been hearing most frequently from clients recently.

Indeed, market concerns are not unfounded. According to Goldman Sachs' latest report data, AI-driven related stocks have risen another 17% year-to-date in 2025, following a 32% surge in 2024. Under such rapid gains, the S&P 500's forward P/E ratio has reached 22 times, placing it at the 96th percentile historically.

However, after thorough analysis, Goldman Sachs provides a relatively measured response: optimistic, but not yet irrational.

The report indicates that current market pricing implies long-term earnings growth expectations of approximately 10%, only slightly above the 9% historical average, but well below the 16% during the 2000 tech bubble and 13% at the 2021 market peak.

Even the market's most prominent "star stocks" - large tech companies - maintain relatively restrained valuations. Report data shows that the top five tech giants by market cap (Nvidia, Microsoft, Apple, Google, Amazon) currently have an average forward P/E ratio of 28 times, significantly below the 2021 peak of 40 times and the tech bubble period's 50 times. Goldman Sachs believes that while current valuations are expensive, they still maintain some distance from the two major historical tops.

AI Trading's "Midfield Battle": Infrastructure Euphoria and Growth Bottlenecks

Goldman Sachs divides the AI trading evolution into different stages in their report, with the current market clearly still immersed in the "second stage" - the euphoria of infrastructure construction.

The fuel for this euphoria comes from hundreds of billions in capital expenditure from major cloud service providers like Amazon, Google, Meta, and Microsoft. Goldman Sachs' report cites data showing that market forecasts for these giants' total 2025 capital expenditure have been revised upward by $100 billion this year alone, reaching $368 billion.

This massive investment directly translates into orders and profits for infrastructure suppliers including semiconductors, power equipment, and technology hardware, driving their stock prices higher.

However, risks lurk beneath the feast. Goldman Sachs warns that the rapid growth in capital expenditure will "inevitably slow down," creating valuation risks for "second stage" stocks. Currently, these companies' stock price gains have far exceeded their recent earnings growth trajectory, reflecting market expectations of extremely optimistic future growth.

Key Risk: Capex Slowdown Could Trigger 15-20% Correction

Goldman Sachs clearly identifies the core vulnerability of current AI trading: dependence on tech giants' capital expenditure. Analysts widely predict that capex growth will show significant deceleration in Q4 2025 and 2026. Once this inflection point appears, it could pressure valuations of related stocks.

While predicting the timing of this inflection point is extremely challenging - as market consensus has repeatedly underestimated tech giants' investment scale - this deceleration trend is "inevitable."

The firm constructed an extreme stress test scenario: if tech giants' capital expenditure suddenly retreated to 2022 levels, this would reduce the S&P 500's expected 2026 sales growth by approximately 30%.

The report's macro valuation model shows this shock would not only directly impact short-term revenue but severely damage market confidence in long-term AI-driven earnings growth, potentially leading to a 15% to 20% downward adjustment in S&P 500 valuation multiples.

What's Next? Market Still Awaits "Earnings Story"

As the infrastructure construction boom gradually peaks, can AI trading's next destination - the "third stage," companies achieving AI-enabled revenue growth - take the baton?

Goldman Sachs' observations reveal market hesitation. The report states that investors show "limited interest" in "third stage" companies, particularly in the software industry.

The reason lies in the market's struggle to answer a complex question: for numerous software and service companies, is AI a growth catalyst or a disruptive threat? Goldman Sachs analysts note that investors worry AI might disrupt existing pricing models, lower industry barriers to entry, and thereby compress profit margins for existing software giants.

Unlike the "second stage" where winners are clearly visible, the "third stage" will see obvious differentiation, with both winners and losers. Therefore, investors have become particularly selective. The report emphasizes that markets may require seeing "tangible, concrete" impact of AI on these companies' near-term earnings before truly embracing them.

As for the longer-term "fourth stage" - AI-driven productivity improvements - Goldman Sachs believes this is still just beginning. The report cites data showing that while 58% of S&P 500 companies mention AI in earnings calls, primarily applied in customer support, programming, and marketing, few companies can quantify AI's specific contribution to current profits.