2025’s biggest investing lesson – slow down

Volatility and the rush to react—not any single policy shock—dominated 2025 investing outcomes: the 2020s have already seen 440 trading days with moves of 1%+ (a typical decade averages ~507). The DeepSeek episode saw Nvidia plunge 17% on claims DeepSeek built a competitive model for US$6m (a figure that omitted prior research costs) before Nvidia reported ~65% revenue and ~57% profit growth YoY and rallied ~60% from its January low. Market timing floundered on rates too—Goldman had forecast five cuts but markets saw three, with an expected March cut (priced ~73%) only arriving in September—and yet the S&P 500 rose over 23% in 2024, underscoring that staying invested through headline-driven volatility has historically outperformed rapid trading strategies.

Analysis

Market structure: The near-term winners are incumbents supplying AI compute (NVDA) and hyperscalers whose multi-year data‑centre capex remains sticky; losers are headline‑driven momentum traders and small AI pureplays lacking customer contracts. Nvidia retains pricing power because software+ecosystem lock‑in raises switching costs; a sustained order backlog implies supply‑constrained pricing for 6–24 months. Cross‑asset: elevated equity volatility lifts option IV (higher hedging costs), pushes some cash into short‑dated Treasuries, and increases USD safe‑haven flows during headline shocks.

Risk assessment: Tail risks include US–China regulatory decoupling (probability ~10–15% next 12 months), surprise supply ramp from cheaper AI accelerators, or systemic liquidity shocks that spike vols >50% (VIX). Immediate (days): 5–15% headline swings; short (weeks–months): earnings/guidance resets; long (1–3 years): secular AI capex growth probable but concentrated. Hidden dependencies: end‑market demand is funneled through few buyers (hyperscalers), so counterparty concentration risk and revenue lumpyness are real.

Trade implications: Prefer concentrated, hedged exposure to NVDA rather than pure momentum plays; use collars or long‑dated OTM calls to capture asymmetric upside while paying down cost with short calls. Rotate into low‑cost broad equity (SPY/VOO) to avoid missing “best 10 days” (target 5–10% portfolio), and maintain a 1% tactical long‑vol sleeve (VIX call spreads 30–90d) to protect against headline crashes. Avoid tariff timing trades; if market pricing implies >50% chance of immediate policy change, reduce cyclical exporters.

Contrarian angles: Consensus that "fastest finger" wins is overdone — volatility creates option‑like opportunities for patient holders. DeepSeek’s low headline cost understates incumbent advantages (ecosystem, IP, silicon scale); the market may underprice durable free cash flow of top AI suppliers. Historical parallel: 2018 tariff/tech headline noise where buy‑and‑hold captured most gains; unintended consequence of crowding is front‑running liquidity squeezes that amplify drawdowns for the fastest traders.

Key Decisions for Investors

• Establish a 2–3% net long position in NVDA by end‑Q1 2026 (size to portfolio), funded from cash or rebalanced cyclical exposure; hedge with a 3–6 month 10% OTM put and sell a 6–12 month 25% OTM call to partially finance the hedge; trim 25% if NVDA quarterly revenue guidance misses by >10% or YOY growth falls below 30%.
• Allocate 5–10% of portfolio to passive US equity (SPY or VOO) by Jan 2026 to capture long‑run compounding and avoid missing the top 10 best days; keep cash allocation ≤10% to reduce timing risk and rebalance monthly.
• Implement a 1% tactical long‑vol hedge: buy VIX 30–60 day call spreads sized to cover expected put costs and rebalance after each volatility spike; deploy when S&P implied vol term‑structure in contango flips to steep backwardation (>+3 vol points).
• Initiate a small 0.5–1% bearish exposure to GS via a 3–6 month put spread (strike width 10–15% below spot) as a pair trade against NVDA long, reflecting risks to trading/forecast credibility; close or reassess within 90 days or after GS reports trading revenue beat >5%.
• Monitor three specific catalysts over the next 60–90 days and act: (1) Nvidia quarterly guidance and capex commentary (add on >15% pullback, trim on >60% upside), (2) US/China AI export or subsidy policy changes (reassess allocations if export curbs enacted), (3) Fed dot‑plot and realized CPI (if rate cut expectations shift by >50bps, rebalance duration and financial exposure).