The 2025 McKinsey Global Payments Report: Key Takeaways

McKinsey's September 2025 Global Payments Report lays bare the forces reshaping how money moves — and why the fraud landscape is about to get significantly more complex. Here are the numbers and themes that matter most for anyone building trust infrastructure.

The Scale of Global Payments

The payments industry remains the most valuable subsector in financial services:

• $2.5 trillion in global payments revenue (2024), projected to reach $3.0 trillion by 2029 at 4% annual growth
• $2.0 quadrillion in total value flows worldwide
• 3.6 trillion transactions processed annually (~9.9 billion per day)
• Average 18.9% return on equity for payments players, with some exceeding 100%

Revenue growth decelerated from 12% in 2023 to 4% in 2024, driven by peaking interest rates, structural shifts toward lower-yield payment methods, and ongoing fee pressures. Interest income made up 46% of total revenues — a figure that will compress as rates decline.

The Cash-to-Digital Shift

Cash usage has fallen to 46% of worldwide payments, down from 50% in 2023. Digital wallets now account for approximately 30% of global point-of-sale volume, led by markets like India (UPI), Brazil (Pix), and Nigeria. Account-to-account (A2A) payments are gaining ground, particularly through instant payment infrastructure.

This shift creates a paradox for fraud prevention: more digital transactions mean a larger attack surface, while lower-yield rails compress the economics available to fund fraud defenses.

Three Forces Reshaping Payments

1. Fragmentation & Regionalization

The era of unified global payment systems is ending. Geopolitical tensions, sanctions, and national sovereignty priorities are driving a proliferation of regional rails. Russia moved to Mir cards after SWIFT exclusion. The ECB is promoting Europe-focused systems. India's UPI is expanding into the Middle East and Southeast Asia.

Fraud implication: Fragmented rails mean fragmented fraud intelligence. Cross-border fraud patterns become harder to detect when transaction data is siloed across incompatible regional systems. This is precisely the gap an orchestration layer like FraudNet is designed to bridge.

2. Stablecoins & Tokenized Money

Stablecoin issuance has doubled since early 2024, with daily transaction volumes around $30 billion. The US GENIUS Act, EU regulations, and frameworks in the UK, Hong Kong, and Japan are providing regulatory clarity. PayPal, Coinbase, and others are already deploying stablecoin-linked payment products.

Fraud implication: Programmable money introduces programmable fraud. Smart contract exploits, stablecoin de-pegging risks, and the "always on" nature of crypto rails (no settlement windows = no cooling-off period) demand real-time, AI-native detection that traditional batch-processing systems cannot provide.

3. AI & Agentic Commerce

This is the force McKinsey flags as most transformative. Key data points:

• 10% of consumers already use AI to start their online shopping journey
• 20% would be comfortable asking AI to make purchases on their behalf
• Visa, Mastercard, PayPal, and Stripe have all launched agentic commerce solutions
• AI agents will "select, optimize, and transact on humans' behalf"

McKinsey calls for "moving intelligence to the edge" — fraud detection, routing logic, and liquidity management embedded directly in software agents and APIs, not centralized in batch systems. This is a direct validation of real-time, API-first trust infrastructure.

The Fraud Prevention Gap

While McKinsey focuses on the payments opportunity, the fraud prevention challenge embedded in these numbers is staggering. Consider the math:

• $2.0 quadrillion in value flows to protect
• 9.9 billion transactions per day to score in real time
• $50 billion in direct payment fraud losses (McKinsey's estimate)
• $10.5 trillion in total cybercrime costs when you include the full stack — operational disruption, remediation, regulatory fines, reputational damage, and downstream economic impact (Cybersecurity Ventures, 2025)

That's a 200× gap between what the industry counts as "fraud" and what organizations actually lose. The three forces McKinsey identifies — fragmentation, programmable money, and agentic AI — will widen this gap unless trust infrastructure evolves at the same pace.

McKinsey's Six Strategies — Through a Fraud Lens

The report concludes with six strategies for payments players. Each one has a direct analog in fraud prevention:

1. Design for intelligent simplicity — Fraud decisioning must be transparent and explainable, not a black box
2. Treat interoperability as infrastructure — Cross-rail, cross-border fraud intelligence sharing is table stakes
3. Move intelligence to the edge — Real-time scoring at the transaction point, not post-hoc batch analysis
4. Make compliance programmable — Modular policy engines that adapt to diverging regional regulations
5. Play through ecosystems — Become the trust layer others build on
6. Earn trust upstream — Transparency and explainability in AI-driven fraud decisions

Bottom Line

McKinsey's report confirms what FraudNet is built for: a world where payments are fragmenting across rails, going programmable, and being mediated by AI agents — all while the fraud threat scales from $50 billion in direct losses to $10.5 trillion in total economic impact. The need for a real-time, cross-vertical orchestration layer that can operate across this complexity isn't a prediction anymore. It's the present reality described by the industry's most authoritative research.