Goldman Clients Take More ‘Cautious Posture’ After November Rout

Goldman Sachs’ Marquee poll of 782 institutional clients shows a notable pullback in bullishness on AI and US equities after November’s selloff, with surveyed investors now expecting the S&P 500 to finish 2026 in a 7,000–7,500 range. That marks a clear retreat from late-October positioning when many clients projected the index near 7,200 by year-end, signaling a more cautious stance that could weigh on risk-taking and portfolio flows into growth/AI-related names.

Analysis

Market structure: Goldman's client de-risking (S&P 2026 median expectation 7,000–7,500 vs prior optimism) favors defensive income and liquidity providers and penalizes crowded, high-multiple AI/mega-cap exposures (NVDA, SOXX, QQQ). Reduced demand for growth beta will lower incremental marginal buyers for passive/ETF product flows, widening bid-ask/friction for large-cap vol and increasing skew. Cross-asset: expect modest bid into U.S. IG/Treasury duration (TLT, LQD) and a temporary rise in equity implied volatility; commodities and FX should see limited directional move absent macro shocks.

Risk assessment: Near term (days-weeks) tail risk is a liquidity squeeze if options/derivatives in largest AI names reprice on concentrated flows; medium term (months) regulatory or earnings disappointments in AI revenue recognition could force de-grossing. Hidden dependency: quant/CTA/prime-broker financing can amplify moves in top-10 names leading to idiosyncratic contagion even if macro is stable. Catalysts to reverse include November-like rebalance, CPI/Fed surprises, or a major AI earnings beat that restores marginal demand.

Trade implications: Favor 1–3% portfolio hedges: buy 3–6 month put spreads on NVDA/QQQ if price drops >15% or 30d IV >60%; allocate 2–3% to long duration (TLT) as convex hedge for next 3–6 months. Implement relative-value pairs: long XLU/XLP vs short XLK/SOXX for 1–3 month tactical rotation; size positions small (1–3% each) given crowding and potential snap-backs.

Contrarian angles: Consensus underestimates persistent enterprise AI spending — a 2026 re-acceleration is plausible, making deep, short-term derisking overdone. Historical parallels (post-2018/2022 tech pullbacks) show rapid mean reversion in concentrated winners once flows normalize; oversized protective trades can be sold into recovery. Unintended consequence: heavy de-risking could create a liquidity vacuum that amplifies rallies into short squeezes in the most-shorted AI names.

Key Decisions for Investors

• Establish a 2–3% portfolio hedge by buying 3–6 month put spreads on NVDA or QQQ sized to limit cost (e.g., buy 15–25% OTM put / sell 35–45% OTM put) if NVDA falls >15% from current levels or 30-day IV >60%; exit on 25–35% realized P&L or 3 months.
• Rotate 2% overweight into TLT or LQD as a convex macro hedge for the next 3–6 months to protect equity risk-off (target duration exposure equivalent to 2–3% portfolio); reduce cyclically exposed tech weight by 2–4% concurrently.
• Implement a 1–2% pair trade: long XLU (or XLP) and short XLK (or SOXX) for 1–3 months; take profits if the sector spread closes by 50% or after 90 days.
• Build a small asymmetric long (1% portfolio) via 9–12 month call spreads on NVDA or SMH (buy 30–50% OTM call, sell 60–80% OTM call) to capture mean reversion in AI capex with limited premium outlay; size modestly to avoid crowding.