Nasdaq poised to lead recovery after tech-led selloff

Tech-led selling pressured the Nasdaq (down 1.6% Thursday) even as futures recovered modestly; Nasdaq futures were +0.6% pre-open while S&P 500 and Dow futures were +0.5% and +0.4% respectively. Alphabet unveiled a $175–$185 billion 2026 capex plan—nearly double last year—and flagged YouTube revenue shortfalls despite 13.5% ad growth and a 48% jump in Google Cloud; Amazon reported sales up 14% to $213 billion with AWS +26% but plunged >10% after-hours on a $200 billion 2026 capex plan ($50 billion above forecasts). The moves have reignited AI-bubble fears, prompted investors to reduce exposure to hyperscalers, and coincided with weakness in crypto and precious metals, amplifying volatility and risk-off positioning.

Analysis

Market structure: The immediate winners are infrastructure suppliers (chips, GPUs, switches), data‑center REITs and exchanges (NDAQ) that capture higher trading/volatility flows; direct losers are hyperscalers (AMZN, GOOGL) facing investor scrutiny over $150–200bn capex plans. Higher capex raises barriers to entry and increases pricing power for chipmakers and equipment vendors, while compressing near‑term FCF for cloud incumbents. Supply/demand signals point to tighter GPU/power/real‑estate availability over 6–18 months, raising input costs and creating single‑sourced vendor leverage.

Risk assessment: Tail risks include AI spending proving non‑revenue generating (operational writeoffs), major chip supply shocks, or regulatory breakups of hyperscalers; any one could trigger multi‑month derisking. Timeline: immediate volatility (days) around earnings/guidance; short term (3–12 months) margin pressure and multiple compression; long term (2–5 years) optionality if AI monetization materializes. Hidden dependencies: enterprise adoption cadence, power pricing, and access to HBM/GPU supply; catalysts are next 2–4 earnings calls, NVIDIA supply updates, and monthly jobs/CPI data.

Trade implications: Tactical plays should be size‑controlled: favor long NDAQ and selective data‑center/semiconductor suppliers, hedge hyperscaler equity risk with targeted puts or pair shorts. Use options to buy downside protection (cost‑limited put spreads) rather than naked shorts given elevated IV; consider harvesting premium in beaten down, but stable, non‑AI names. Rotate 5–10% of equity exposure from large cap hyperscalers into infrastructure/REITs over 4–12 weeks.

Contrarian angles: Consensus misses that front‑loaded capex can create 2027–2029 revenue leverage — if AI workloads scale, hyperscaler revenues could re‑accelerate and current multiple compression would reverse. The market may be overpricing permanent impairment risk; history (previous cloud cycles) shows capex spikes often precede outsized revenue growth 18–36 months out. Unintended consequence: heavy capex benefits suppliers and smaller cloud challengers who lease capacity, creating asymmetric opportunities for long suppliers vs. short hyperscalers.

Key Decisions for Investors

• Establish a 1.5% portfolio long in NDAQ (Nasdaq) over 6–12 months to capture higher trading/volatility flow; set a 10% stop-loss and add to position on any >8% pullback.
• Implement a pair trade: go long 1.5% NDAQ and short 1.5% AMZN (equal notional) to express infrastructure upside vs. hyperscaler capex risk; rebalance after earnings season (next 6–8 weeks).
• Buy a cost‑limited AMZN downside hedge: purchase a 6‑month put spread (buy 25% OTM, sell 12% OTM) sized to protect 1–2% of portfolio; increase protection to 3% if AMZN falls an additional 10% from current levels.
• Reallocate 3–5% from FAANG/hyperscaler exposure into data‑center REITs (e.g., EQIX) or top semiconductor suppliers over 3 months, targeting 6–18% IRR if AI capex drives leasing and chip demand; trim if REIT yields compress by >150bp.
• If you hold >5% position in GOOGL/AMZN, reduce to 3–4% within 30 days and replace with infrastructure longs unless next earnings show sustained revenue acceleration or capex ROI metrics (unit economics) improve within two quarterly updates.