American Express Is Winning the Gen Z Premium Card War — And Betting $5B a Year on Tech to Keep It

Amex posted record FY2025 revenue of $72.2B, grew net card fees 18% to $10B, and raised 2026 guidance above consensus. The real story: a $5B annual tech spend, an aggressive Gen AI rollout across 70+ use cases, and a Platinum Card refresh that's pulling younger demographics at rates that would make fintech founders jealous.

AXP · Financials · February 18, 2026

S&P 500 Position

Amex sits in the Financials sector (~12.8% of S&P 500 weight), alongside Berkshire Hathaway, JPMorgan, Visa, and Mastercard. It's smaller than Visa (~$600B) and Mastercard (~$450B) by market cap but carries more credit risk and earns higher margins per transaction due to its closed-loop model. Within Consumer Finance specifically, it dominates the premium tier while Capital One and Discover compete on mass-market and subprime segments. The pending Capital One–Discover merger could create a new closed-loop competitor.

Index Weight: ~0.46% | Rank: Approximately #40-50 in the S&P 500 by market cap

Company Overview

American Express is executing one of the most disciplined premium-customer growth strategies in financial services. Its closed-loop network — where Amex simultaneously issues cards, acquires merchants, and runs the payment rail — gives it a data advantage that pure-play processors like Visa and Mastercard cannot replicate. The company captures both ends of every transaction, enabling tighter fraud models, richer personalization, and a direct customer relationship that powers its $10B net card fee business. In September 2025, the company rolled out its biggest-ever Platinum Card refresh, raising the annual fee to $895 while layering on over $3,500 in annual credits spanning Resy dining, Fine Hotels + Resorts, lululemon, Oura Ring, and Uber — a clear play for millennial and Gen Z customers living in major metros. The generational bet is paying off. Gen Z spending on U.S. Consumer cards surged 38% year-over-year in 2025, and 73% of global new accounts came from fee-paying products. Amex spent $6.3B on marketing and $5B on technology in 2025, both up meaningfully. The company is migrating its data and analytics infrastructure to a third-generation platform targeted for completion by 2027, which has already cut processing times for marketing and fraud workflows by 90%. With Squeri publicly naming agentic commerce as the next frontier, Amex is positioning itself not just as a payment network but as the trusted identity and payment layer for AI-agent-driven transactions. The FY2025 numbers tell a clean story: $72.2B in revenue (up 10%), $15.38 EPS (up 15% adjusted), 34% ROE, and net write-offs holding at 2% — effectively at the floor for consumer lending. The 2026 guidance of 9-10% revenue growth and $17.30-$17.90 EPS signals the company sees no meaningful slowdown in premium consumer spending.

Products & Revenue

Amex's revenue engine is built on three pillars: discount revenue (the merchant fee collected on every card swipe, driven by $1.67T in billed business), net card fees ($10B in FY2025, growing 18% YoY as premium products scale), and net interest income ($15.5B in FY2024, surging as revolving loan balances grow among a creditworthy base). The company also generates revenue through travel services, insurance, and other fees. Unlike Visa/Mastercard, Amex takes on credit risk directly, which means higher margins when credit quality holds — and it has, with write-off rates at 2%.

U.S. Consumer Services (USCS) (~48%): Issues premium consumer cards (Platinum, Gold, Green, Blue) in the U.S., operates the high-yield savings account, and runs Amex Travel/Resy. The Platinum refresh and Gen Z acquisition engine live here.

Commercial Services (CS) (~23%): Provides small business, mid-market, and large corporate charge/credit cards, plus expense management tools and working capital solutions. Includes the Business Platinum and Kabbage-derived lending products.

International Card Services (ICS) (~18%): Consumer and small business card issuance outside the U.S. Billed business grew 12% FX-adjusted in Q4 2025, the fastest-growing card segment. Operates in ~130 countries.

Global Merchant and Network Services (GMNS) (~11%): Runs the Amex payment network, third-party card issuing partnerships (GNS), merchant acquiring, and network services. This is the toll booth of the closed-loop model.

Based on FY2025 segment data from tickergate.com and Q4 2025 earnings release (January 30, 2026). Percentages are approximate.

Leadership

Stephen J. Squeri

CEO since 2018. A 40-year Amex veteran who rose through technology and operations, including a stint as CIO from 2005-2009. Squeri succeeded Kenneth Chenault and has driven the company's pivot toward premium millennial/Gen Z acquisition, the $5B annual tech investment strategy, and the push into agentic commerce. Named to TIME's 100 Most Influential People in 2025.

Christophe Le Caillec, Chief Financial Officer: French-born, joined Amex in 1997 and has held global roles across Paris, Sydney, Singapore, and London. Previously served as Deputy CFO and CFO for Global Consumer Business. Key architect of the capital allocation strategy that returned $5.3B in buybacks and raised the dividend 16% for 2026.

Anré Williams, CEO, American Express National Bank & Group President, Enterprise Services: Wharton MBA, Stanford BA. Led Global Merchant Services before taking on enterprise-wide responsibilities. Oversees key operational functions and sits on the Executive Committee driving strategic direction.

Howard Grosfield, Group President, U.S. Consumer Services: Leads the largest revenue segment and drove the 2025 Platinum Card refresh. Publicly cited Amex's 98% retention rate and the strategy of acquiring Gen Z customers early to capture 20+ years of lifetime value.

Mohammed Badi, President, Global Network Services: Former Chief Strategy Officer, now leads network-related businesses including payment network operations, global network partnerships, and emerging payment products — the infrastructure layer of the closed-loop model.

The AI Angle

From fraud ML to agentic commerce payment rails

American Express has been deploying machine learning since 2010, when it first applied ML to fraud detection. Today, the company is exploring more than 70 generative AI use cases across the organization. The most visible consumer-facing deployment is the Gen AI-powered Travel Counselor Assist tool, which gives human travel agents real-time recommendations for complex requests — booking Maldives itineraries via Dubai with vegan dining constraints, for example — reducing hold times and eliminating callback cycles. The Amex app has also integrated Gen AI-powered experiences referenced in the Q4 2025 earnings commentary. On the corporate side, Amex Global Business Travel has shipped LLM-powered virtual agents that handle complex travel modifications through natural language, plus a Natural Language Query feature in Egencia Analytics Studio that lets travel managers ask questions like 'what is spend, transaction count, ATP for air in city Paris 2024?' Infrastructurally, Amex partners with NVIDIA for fraud detection models and is migrating its entire data and analytics stack to a third-generation platform targeted for 2027 completion. Early results have cut processing times for marketing and fraud workflows by 90%. Within Credit and Fraud Risk, the company created a Frontier Research Team — a dedicated internal research unit tasked with long-term AI/ML investigation beyond production model maintenance, focused on methodologies that could be transformational in credit and risk decisioning. On the venture and ecosystem front, Amex Ventures has launched a dedicated Gen AI investment track and is actively scouting startups. A key strategic bet is the $5M seed investment in Nekuda, which is building infrastructure for AI agents to make autonomous payments — what the industry calls 'agentic commerce.' Squeri has publicly called agentic commerce the next big trend in payments, stating AI will help consumers find the best deals and choose the optimal payment method automatically. Amex is directly shaping EMV specifications through EMVCo to create standards for secure card-based agentic payments. The competitive positioning is strong: Amex's closed-loop network gives it richer data than open-loop processors. Where Visa and Mastercard optimize network-level transaction flows, Amex knows both the cardholder and the merchant — enabling deeper personalization models and more precise fraud scoring. The risk is execution speed; if agentic commerce scales before Amex's rails are ready, the company could cede ground to fintech-native solutions building AI-first payment infrastructure.

Financial Snapshot

Revenue (TTM): $80.5B — TTM (ending Dec 2025) | Net Income: $10.8B net income

Margins: Net margin 13.5%; operating leverage improving as tech and automation reduce cost-to-serve

Amex's 34% ROE on a 13.5% net margin is the signature of its model: high-spending, low-default customers producing dense fee revenue on top of interest income. The company returned $5.3B via buybacks in 2025, reducing share count 7% since 2022, while raising the dividend 16% for 2026 to $0.95/quarter. The 2026 guidance of $17.30-$17.90 EPS implies continued mid-teens earnings growth, supported by the Platinum fee increase hitting renewals and net card fee momentum showing no signs of deceleration.

1-Year Performance

$344.53 current price. The stock's 52-week range spans $220.43 to $387.49, meaning it sits roughly in the upper third of its annual range.

AXP has outperformed the S&P 500 every year since 2021 — remarkable for a financial stock during a period dominated by tech. The April 2025 low of ~$220 reflected broad macro fears and tariff concerns, but the stock recovered sharply as premium consumer spending proved resilient. The December 2025 all-time high near $384 was driven by Q3 results and the Platinum refresh momentum. Post-Q4 earnings, shares dipped ~3% on a penny EPS miss despite a strong 2026 guide, indicating the market is pricing for perfection.

Recent News

Fun Fact: American Express began using machine learning for fraud detection in 2010 — four years before Google acquired DeepMind. The company's fraud models now process every single transaction in real-time across its network, and its closed-loop data advantage means it trains on both cardholder behavior patterns and merchant-side signals simultaneously, a dual-sided dataset that open-loop networks like Visa and Mastercard structurally cannot replicate. Internally, the company created a 'Frontier Research Team' within its Credit and Fraud Risk division — modeled after corporate research labs — specifically to give data scientists time to pursue long-horizon AI research without production model maintenance deadlines.