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

Alphabet reported Q4 2025 revenue of $113.8B (+18% YoY), net income +30% YoY and EPS $2.82 (+31%), with Google Cloud revenue up 48% to $17.7B; management plans $175–185B capex in 2026 and holds ~$127B in cash/marketable securities, valuing the stock at ~28x earnings. American Express reported Q4 2025 revenue of $19.0B (+10% YoY) and a record $72.2B for the year, with $10B in net card fees, and represents >$45B (~15%) of Berkshire’s equity portfolio; its P/E is ~20 but is constrained by liability-sensitive lending exposure. The analyst prefers Alphabet as the better buy given steeper, broad-based growth and expanding cloud margins, while noting risks from Alphabet’s heavy AI capex and American Express’s credit-cycle sensitivity.

Analysis

Berkshire’s simultaneous ownership of a franchise banked on closed-loop consumer spend and a fast-scaling AI/cloud platform creates an asymmetric portfolio exposure: one leg benefits from cyclical consumer credit carry, the other from secular scale in high-margin software and AI infrastructure. The second-order winners from Alphabet’s capex push are not only GPU suppliers but also data-center landlords, power equipment vendors and specialized cooling/optics suppliers — expect tighter supplier lead times and pricing power in those niches over the next 12–24 months.

American Express’s core advantage — a premium, fee-dense customer base — also makes it more exposed to macro credit shocks than pure network operators; a shallow recession would compress loan book returns and force higher provision cycles even as fee income proves sticky. Conversely, Alphabet’s risk is executional: multi-year capex can create negative free-cash-flow phases and leaves management accountable to a long-looking IRR threshold on AI infrastructure; failure to convert incremental capacity into higher margin enterprise ARR would quickly re-rate the multiple.

Near-term catalysts to watch are (1) macro credit indicators (delinquency curves, unemployment claims) that presage AXP provisioning, (2) quarterly cloud incremental margins as the clean signal of AI monetization, and (3) GPU supply/pricing and capex cadence that drive vendor earnings. Governance and allocation differences matter: Berkshire’s historical patience on AXP suggests downside support in extreme stress, while Alphabet’s public-market discipline will be judged by FCF recovery post-capex — timeline 12–36 months for clear outcomes.

Key Decisions for Investors

• Pair trade (12–24 months): Long GOOGL (equal-dollar) / Short AXP — target spread appreciation of 30–50% if cloud margins expand while consumer credit weakens. Size: 2–4% net notional; stop if pair P&L down 15% or if macro indicators (3-month rolling unemployment) improve materially.
• Long GOOGL LEAPS (18-month): Buy Jan/Jul 2028 calls ~1–2 strikes OTM to capture AI monetization with limited downside. Position size 1–2% NAV; risk/reward ~payoff skewed: 30–60%+ upside if cloud margins and ad resilience persist, max loss = premium paid.
• Long NVDA (6–12 months) or targeted data-center vendors to play spillover from Alphabet capex — thesis: GPU/contention and retrofit demand sustain pricing. Trade size 1–3% NAV; take profits on 40–60% move or if GPU spot pricing falls >30%.
• Defensive allocation: rotate some payments exposure from AXP into V/MA (6–12 months) — preserves payments upside without lender balance-sheet cyclicality. Use 3–5% reallocation; stop-loss 12% and reassess on improved credit trends.
• Risk hedge (12 months): Buy protection on AXP (buy 1–2x put spread) funded by selling covered calls on V/MA to limit drawdown from an idiosyncratic credit stress event. Cost neutral to modest debit; downside mitigation if consumer credit deteriorates sharply.