Only 3 of My Top 10 Stocks for 2025 Lost to the Market. Here Are the Ones I Think Are Still Buys for 2026.

Meta, Amazon and PayPal all underperformed the S&P 500 (which rose 16.4% in 2025) for different reasons: Meta was punished after guiding higher data-center capital expenditures for 2026, Amazon's valuation premium largely unwound despite improving recent results, and PayPal remains a slow-growing turnaround trading at under 10x forward earnings. The author argues all three could outperform in 2026 if Meta’s data-center investments generate strong returns, Amazon sustains its recent business strength, and PayPal accelerates EPS growth and continues buybacks, but confidence is lowest for PayPal.

Analysis

Market structure: The immediate winners are Amazon (AMZN) and semiconductor/infra suppliers (ASML, TSM) because renewed data‑center and e‑commerce strength drives supplier capex and AWS margin upside; losers are Meta (META) near‑term due to higher disclosed 2026 data‑center capex and PayPal (PYPL) for underdelivering growth. Pricing power shifts toward cloud incumbents and capital‑goods vendors; advertisers may reallocate budget if META’s ROI from incremental compute is delayed, pressuring ad CPMs across the social stack.

Risk assessment: Tail risks include regulatory shocks to payments (10–15% 12–24m probability) or a tech‑capital‑spending bust if AI ROI disappoints (15–20%); macro recession would compress ad/TPV volumes and widen credit spreads. Time horizons: days—sentiment swings on guidance; weeks–months—earnings and Prime/holiday cadence; 12–36 months—realization of data‑center ROI. Hidden dependencies: META’s margin thesis depends on model inference density, wholesale power costs, and silicon availability; PYPL depends on TPV mix, take‑rate expansion and buyback cadence.

Trade implications: Tilt overweight AMZN and ASML/TSM, tactical underweight META until ROI evidence; use pair trades (long AMZN / short META) to express conviction while hedging market beta. Options: buy limited‑risk PYPL Jan‑2027 call spreads to capture a low‑probability, high‑upside turnaround; buy short‑dated protective puts on large tech longs around earnings windows. Reallocate 3–6% from cyclical consumer into tech infra over 3–12 months.

Contrarian angles: Consensus penalizes capex reflexively—historic analog: Amazon’s early heavy capex preceded durable moat expansion; META could be similarly mispriced if incremental compute yields >10–15% ad margin lift. PYPL’s sub‑10x forward PE implies the market demands ~10–15% EPS CAGR or sustained buybacks to re‑rate; mispricing risk exists but requires event‑driven proof. Watch for unintended consequences: an overcommitment to capex could amplify supply‑chain demand for ASML/TSM shares, pushing those equities into momentum froth and creating shortable ripples.

Key Decisions for Investors

• Establish a 2–3% long position in AMZN (shares) within 2–6 weeks, target +20–30% in 6–12 months if revenue and operating margin beat consensus by >=150bps; place an initial stop‑loss at -10% under entry and trim half on +20% gain.
• Initiate a dollar‑neutral pair trade: long AMZN (1.5%) / short META (1.5%) for a 3‑month tactical window; unwind if META cuts 2026 capex guidance or if AMZN misses revenue by >1.5% or margin by >100bps on a quarter.
• Take a contrarian small position in PYPL: 1% long equity plus a Jan‑2027 call spread 25–35% OTM (debit spread) sized to limit max loss to the equity allocation; objective is >2x payoff if PYPL delivers >=10% EPS CAGR or resumes aggressive buybacks within 12–24 months. Exit/trade down to zero if management suspends buybacks or TPV growth falls YoY for two consecutive quarters.
• Add 0.5–1% exposure to ASML or TSM as a beneficiary of incremental data‑center capex; use a 12% stop‑loss and scale out 50% on a +20% price move within 6–18 months.
• Purchase 1–1.5% of portfolio notional in 2–3 month protective puts (or buy‑write collars) on concentrated tech longs ahead of quarterly earnings (next 30–60 days) to cap downside risk if sentiment reverts.