Alphabet vs. American Express: Which Warren Buffett Stock Is a Better Buy?

Alphabet reported Q4 2025 revenue of $113.8B (+18% YoY), Google Services $95.9B (+14%), Google Cloud $17.7B (+48%), with net income +30% and EPS $2.82 (+31%); management plans ~$175–185B capex in 2026 to meet AI compute demand. American Express, Berkshire’s ~15% equity stake (~$45B position), posted Q4 2025 revenue of $19.0B (+10%) and record FY revenue $72.2B, with $10B in net card fees and 30 consecutive quarters of double‑digit card fee growth, but carries liability-sensitive credit risk. The author prefers Alphabet as the better buy due to steeper, broad-based growth, expanding cloud profitability and a strong, cash-heavy balance sheet despite a higher P/E (~28 vs ~20 for AmEx) and elevated capex risk.

Analysis

Alphabet’s growth lever is not just top-line acceleration but an operational flywheel: aggressive hyperscale capex plus proprietary model/infra investments can tighten margin dispersion in cloud services and force competitors to either match spend or cede pricing. That dynamic creates a multi-year winner-takes-most outcome for core infra suppliers and software vendors that can internalize AI compute, while pressuring smaller cloud/SaaS players’ margins as they absorb higher unit compute costs.

American Express sits on the opposite trade-off: franchise durability and fee-rich customer economics come with balance-sheet sensitivity to consumer credit cycles and macro shocks. That makes AmEx a convex credit-beta instrument — attractive if delinquencies keep falling, hazardous if late-cycle unemployment/inflation shocks reappear — and creates asymmetric outcomes versus a capital-light ad/cloud compounder like Alphabet.

Near-term catalysts that will re-rate either name are clear and different in cadence: advertising/enterprise cloud demand and quarterly capex commentary will move Alphabet within days–months, while consumer credit flows, co-brand portfolios, and loan-loss reserve adjustments will move AmEx more over quarters. The consensus tilt toward Alphabet ignores the multi-year cash-cycle drag from hyper-scale capex should AI compute economics improve faster than expected; conversely, AmEx may be under-owned if the market overestimates a severe credit drawdown that never materializes.

Key Decisions for Investors

• Pair trade (12–24 months): long GOOGL equal-dollar vs short AXP 50% notional. Rationale: capture divergence between capital-light, secular cloud upside and balance-sheet cyclicality. Risk management: trim longs on a 20% drawdown from entry or if Google’s quarterly cloud margin growth stalls for two consecutive quarters; size so max portfolio drawdown ≤3%.
• Directional options (18–30 months): buy 18–30 month GOOGL call spreads 20–30% OTM funded by selling 3–6 month calls 5–10% OTM. Expected payoff: asymmetric upside participation if cloud margins expand with capped cost; max loss = net premium (target R:R 3:1 if core thesis realized).
• Tail protection for AmEx (6–12 months): buy AXP puts ~10–15% OTM as insurance (small position). Use this when macro leading indicators (initial jobless claims, payroll surprises) flash deterioration; cost is the premium but protects against a credit event that would knock 20–30% off price.
• Contrarian accumulation trigger for AXP (12–18 months): initiate staged buys if 2 consecutive months of improving card delinquency/travel spend and stable reserve builds; hedge with short-term puts to keep net exposure neutral until two quarterly prints confirm normalization. Target return if cycle stabilizes: 20–40% with hedges limiting downside to ~10–12%.