Q1 2026 Results vs. Consensus

Thesis Pillar Assessment — Post Q1 2026

Pillar 1: Valuation Premium Unsustainable — ✅ ON TRACK

Status: The print has not rescued the valuation case for bulls. At $320.51, AXP trades at 18.2x our 2026E EPS of $17.56 — still 7–14% above our 16–17x mean-reversion target. Trailing P/B remains ~6.4x, implying a perpetual ROE above 48% at our cost of equity of 10.62%. Our FCFE / Residual Income base case is unchanged at $266 per share.

Multiple Impact: The 3.5% post-earnings sell-off shows investors are beginning to question the valuation ceiling. With guidance merely reaffirmed (and management signalling increased investment spend), there is no near-term EPS catalyst to justify multiple expansion beyond current levels. A beat-and-maintain is incrementally de-rating, not re-rating.

Conviction: Moderate-High — Unchanged. Valuation remains the most durable pillar of our SELL thesis.

Pillar 2: Late-Cycle Credit Normalization — ⚠️ DELAYED (Conviction Reduced)

Status: The Q1 data is running counter to our thesis in the near term. The principal-only net write-off rate came in at 2.0% — below 2019 pre-pandemic levels and below our FY25 initiation assumption of 2.3%. Management guided 'generally stable credit metrics throughout 2026,' effectively ruling out any near-term normalisation toward our 3.0% thesis trigger.

NCO Trajectory vs Thesis: Our initiation modelled NCO rising from 2.3% (FY25) toward 3.0–3.2% by FY27 as late-cycle pressures build. Q1 2026 shows 2.0% — a 30bp downward miss. If AXP maintains sub-2.2% NCO through FY26, provision tailwinds add ~$0.30–0.50 to EPS vs our model, which would support current valuation for longer.

Reserve Release: $24M reserve release in Q1 is consistent with a consumer credit environment that is benign by any historical standard. This is a cyclical phenomenon driven by resilient upper-income consumer spending, not a structural improvement in AXP’s underwriting.

Conviction: Reduced from Moderate → Moderate-Low. The thesis is not disproven — credit cycles turn when they turn. But we are clearly in an extended benign phase, and Pillar 2 is not a near-term catalyst for our SELL.

Pillar 3: Premium-Card Competitive Intensification — ✅ ON TRACK (Mixed Signals)

Status: The rewards cost pressure we identified at initiation is materialising. Rewards expense rose $513M year-over-year in Q1 (Membership Rewards + cash-back +$353M; cobrand +$160M). Variable customer engagement costs reached 44.7% of revenue, up ~70bps YoY — directionally consistent with our thesis of rising rewards burden squeezing operating leverage.

Offsetting Factor: Net card fees grew 18% YoY, reflecting successful fee hikes on the Platinum and Gold cards. This is the one element where AXP has demonstrated pricing power that is suppressing the competitive-intensification signal — card-fee revenue is partially absorbing the rewards cost escalation.

Rewards / Discount Revenue Ratio: Not separately disclosed in Q1 press materials. Variable customer engagement at 44.7% of total revenue captures rewards + cardmember services + co-brand costs holistically. The directional trend (rising) is consistent with our FY24 benchmark of 53% rewards/discount revenue rising toward 55–56%.

Conviction: Moderate — Unchanged. Rewards inflation is confirmed; net card fee resilience delays full margin compression.

Pillar 4: Disruptive Threat Optionality — ⚠️ PARTIALLY MITIGATED (Significant New Development)

Status: This is the morning note’s most strategically important update. AXP is no longer a passive target of AI-agent commerce disruption. The company is actively embedding itself as the infrastructure layer for AI-mediated transactions:

• Agentic Commerce Experiences developer kit: enables third-party AI agents to initiate, authenticate, and complete AXP-backed transactions programmatically.
• Amex Agent Purchase Protection: extends card-member protections to AI-initiated purchases — a key trust barrier for AI-agent adoption.
• Hypercard acquisition: AI-focused expense platform acquisition extends closed-loop network into intent-based, AI-driven corporate spending.

Our Take: AXP’s strategy is defensible: if AI agents become the dominant commerce interface, AXP wants to be the trusted payment rail, not the disrupted incumbent. The closed-loop network — where AXP controls both issuing and acquiring — is a genuine structural advantage for embedding agent-level controls. However, the risk is not fully mitigated: (1) open-loop rails (Visa/Mastercard) and BNPL players are also building agent-commerce integrations; (2) stablecoins and RTP threaten discount-revenue economics regardless of AXP’s AI moves; (3) the Hypercard acquisition is small and unproven at scale.

BNPL/RTP: No material update from management. Both continue to grow sub-threshold for our trigger events.

Conviction: Reduced from Moderate-Low → Low. AXP’s proactive AI positioning is a genuine mitigation. Pillar 4 remains part of the thesis as asymmetric downside optionality, but the imminent disruption probability has declined.

Disruptive Threat Register — Updated May 4, 2026

Management Commentary — Key Themes

Guidance Reaffirmation: Reading Between the Lines

Management reaffirmed FY2026 guidance of 9–10% revenue growth and EPS $17.30–$17.90. The key sentence: the overperformance in Q1 will be reinvested into marketing and technology rather than flowing through to the bottom line. This signals two things: (1) competitive intensity is real — AXP sees a need to invest aggressively to defend market share; (2) management is deliberately front-loading investment, suggesting confidence in the top-line trajectory but skepticism about pure operating leverage flow-through.

Capital Return

Average diluted shares declined 2% YoY to ~686M, confirming active buyback execution. No specific quantum was disclosed in press materials. The active repurchase programme, combined with an 18% card fee growth and a maintained dividend, reflects a capital-allocation approach that is somewhat inconsistent with a company that — at 6.4x P/B — is buying back its own stock at a 48%+ implied perpetual ROE requirement. We view this as value-destructive at current prices.

Late-Fee Rule (CFPB)

No new commentary from management on the CFPB $8 late-fee cap. The rule remains in litigation limbo following district court injunctions. AXP’s exposure is structurally lower than mass-market issuers (premium card-members are less delinquent), but the industry-wide precedent poses revenue risk. We monitor for court outcomes in H2 2026.

T&E Momentum

Travel & Entertainment spending grew 12% (9% FX-adjusted), reflecting strong restaurant performance and airline spend. However, management noted airline spend softened in the final weeks of Q1 due to travel disruptions from the Middle East conflict. This is a macro risk to monitor — T&E is the highest-margin spend category and the primary justification for premium-card pricing.

Post-Print Stock Reaction

Market Interpretation: The post-print pattern is instructive. AXP beat on every line, yet the stock fell. This is the archetypal ‘priced for perfection’ dynamic — a beat-and-maintain is not sufficient to justify a 18.2x forward multiple. The market is starting to question whether guidance conservatism (a traditional AXP management trait) is hiding structural growth deceleration or whether the reinvestment commentary signals margin compression ahead.

Today’s modest stabilisation (+0.6%) reflects short-covering and value-seeking at the $319-320 support level. We do not view this as a trend reversal.

Thesis Status — SELL Maintained

Upcoming Catalysts

Sources: American Express Q1 2026 Earnings Press Release (Apr 23, 2026); AXP Q1 2026 Earnings Call Transcript; GuruFocus, Grafa, IndexBox, Simply Wall St, Investing.com — Q1 2026 earnings coverage; AgenticFinance Chile Initiation Report (Apr 30, 2026).

NOTE ON EARNINGS DATE: AXP reported Q1 2026 results on April 23, 2026 (not May 1 as originally briefed). All financial data reflects actual reported figures from that date. May 4 note issued as first Monday morning following our coverage initiation and the post-print settlement period.

This note is produced by AgenticFinance Chile for informational purposes only. Not investment advice. All data sourced from public filings and third-party research. Past performance is not indicative of future results.