[Goldman Sachs Report] The True Driver of the S&P 500 is AI, Not Oil: Earnings Trajectories and the Two Faces of Investment

[Goldman Sachs Report] The True Driver of the S&P 500 is AI, Not Oil: Earnings Trajectories and the Two Faces of Investment

Subtitle: Analyzing the Limits of Short-Term Oil Shocks and the Surging Depreciation Burden on Hyperscalers

Original Report Date: March 6, 2026

---

3-Line Summary

• Limited Impact of Oil Shocks: An analysis of 7 major geopolitical crises since 1950 shows that equity market hits are short-lived (averaging a 4% drop followed by recovery within a month). The impact of surging oil prices on overall S&P 500 earnings is also largely neutral, as effects cancel out across different sectors.

• Structural Shift & AI Dominance: The energy sector's earnings weight in the S&P 500 has plummeted to 4%, while the IT and Communication Services sectors have surged to 38%. Notably, Nvidia alone is projected to drive 24% of the entire S&P 500's earnings growth this year (2026).

• The AI Investment Dilemma (D&A and FCF): While massive Capital Expenditure (Capex) by Big Tech heavily benefits the AI infrastructure value chain, hyperscalers face a surging Depreciation & Amortization (D&A) to sales ratio (expected to hit 14% by 2027). This, combined with deteriorating Free Cash Flow (FCF), is significantly adding to valuation burdens (Forward P/FCF at 115x).

---

In-Depth Report Analysis

With recent news of war in the Middle East and surging oil prices, stock market volatility has increased. However, Goldman Sachs (GS) diagnosed that the direct impact of geopolitical risks and rising oil prices on US equities may be more limited than thought. Instead, they emphasized that the true core variable dictating the medium-to-long-term direction of the S&P 500 is the "trajectory of AI infrastructure investment and its monetization." Let's examine in detail why GS is paying more attention to the "depreciation" of hyperscalers than to macroeconomic indicators like oil.

1. Geopolitical Risks and Surging Oil Prices: Just a Passing Shower?

Last week, the S&P 500 dropped 2% on oil spike fears. However, Goldman Sachs cautioned against excessive market panic, citing historical data.

Analyzing 7 major geopolitical risks since 1950, GS found that the S&P 500 suffered an average decline of 4% in the first week but showed resilience by recovering to previous levels within a month in most cases. Furthermore, the overall impact of rising oil prices on corporate Earnings Per Share (EPS) is evaluated as 'neutral'. The boost in earnings for energy companies is offset by the deteriorating performance of sectors sensitive to raw material costs, such as airlines and consumer discretionary goods.

The exception, of course, is if crude oil supply disruptions prolong and cause a slowdown in actual real economic growth. According to GS models, every 1 percentage point drop in US real GDP growth reduces S&P 500 earnings by 3-4%. Thus, it is far more critical to monitor whether the oil price hike actively suppresses consumer sentiment and triggers actual 'economic contraction' rather than focusing on the price hike itself.

2. S&P 500 Structural Shift: The Exit of Energy, the Dominance of AI

The fundamental reason Goldman Sachs remains relatively calm about the oil issue lies in the profound structural changes in the US stock market.

In the 1980s, the energy sector accounted for a massive 30% of total S&P 500 earnings; today, it has shrunk to a mere 4%. Even when combining all consumer discretionary sectors sensitive to oil, the weight only reaches 18%. In stark contrast, the IT (25%) and Communication Services (13%) sectors now account for 38%, proving that the market's center of gravity has completely shifted to tech, specifically the 'AI' theme.

This impact is highly visible in earnings numbers. While average earnings for S&P 500 companies grew by 9% in 2025, the earnings of the 7 major AI-related stocks (Magnificent 7) skyrocketed by 32%, accounting for half of the index's total earnings growth. Furthermore, GS projects that Nvidia alone will generate 24% of the total S&P 500 earnings growth this year (2026). GS estimates that the contribution of AI infrastructure and cloud services to overall S&P 500 earnings growth will expand significantly from approximately 25% in 2025 to 40% in 2026.

3. The Light and Shadow of AI Investment: Surging D&A and FCF Pressures

The epicenter of these massive AI profits lies in the astronomical Capital Expenditures (Capex) by hyperscalers building out massive cloud infrastructure. Their Capex is projected to grow explosively from $412 billion in 2025 to $667 billion in 2026. This capital directly translates into earnings rallies for semiconductor infrastructure companies.

However, GS points out a critical headwind the market is overlooking: the surge in 'Depreciation & Amortization' (D&A) among hyperscalers. As Big Tech heavily procures expensive AI chips, the amortization costs that must be recognized on their balance sheets over the coming years are snowballing. Over the past decade, their D&A-to-sales ratio averaged 8%, but it is expected to spike to 14% by 2027. Naturally, as the annual expenses to be deducted increase, net profit margin forecasts will inevitably be compromised.

An even more critical metric is 'Free Cash Flow' (FCF). Over the past 12 months, the book earnings of hyperscalers increased by 23%, but due to massive upfront infrastructure costs, their FCF—the actual cash left in the corporate vault—contracted by -32%. As a result, their Forward Price-to-FCF (P/FCF) ratio has soared to a historic high of 115x. Due to the time lag between massive capital outlay and actual profit recovery, Big Tech companies are highly likely to face prolonged overvaluation debates.

4. The True Game Changer: The Current State of 'Productivity Gains'

For the market to overcome the burden of these infrastructure costs and leap to the next level, substantial 'productivity gains' (cost reduction and profit maximization) through AI adoption must be proven across all industries.

However, looking at the data objectively, GS diagnoses that we are still in the extremely early stages. While 70% of S&P 500 companies mentioned AI during the recent earnings season, only 1% disclosed quantified, specific cost savings figures from AI adoption. Converted to monetary value, this amounts to roughly $300 million—less than 0.1% of the total projected S&P 500 earnings for 2025.

The silver lining is that AI-related revenue for cloud service providers is growing explosively at over 100% annually, albeit from a small base. Based on this, GS forecasts that the contribution of AI-driven productivity gains to S&P 500 earnings growth will gradually and clearly expand from 0.4 percentage points this year (2026) to 1.5 percentage points next year.

5. Conclusion: Maintaining the 7,600 Target

In conclusion, despite recent geopolitical noise and market volatility, Goldman Sachs expressed confidence in fundamentals. The Earnings Per Share (EPS) of S&P 500 companies is expected to achieve solid 12% year-over-year growth this year. Furthermore, GS announced it firmly maintains its year-end S&P 500 index target of 7,600 (representing roughly an 11% upside from the current 6,800 levels).

---

StockHub Insight & Comments

When the stock market shakes at the sound of artillery in the Middle East, Goldman Sachs quietly looks into the deepest parts of financial statements: 'Depreciation & Amortization (D&A)' and 'Free Cash Flow (FCF)'. The insight this report provides is clear. Short-term oil fluctuations cannot break the intrinsic fundamentals of the S&P 500, but the billing costs (depreciation) stemming from Big Tech's over-investment in AI infrastructure have begun to exert real pressure on valuations.

While cheering for Nvidia's astonishing earnings growth, investors must keep a close eye on the deteriorating cash flows of hyperscalers—such as Microsoft, Google, Meta, and Amazon—who are buying up those chips. The core narrative of the upcoming market phase will shift from "Who built the best AI infrastructure?" to "Who is the first to generate actual revenue (productivity gains) from that infrastructure to cover book costs with real cash?"