Cardinity is now part of
Most payment predictions read like a shopping list. AI this, blockchain that. Nobody stops to explain why these things matter right now, or what they actually mean for the people building and buying a payment infrastructure.
So here’s what’s different about 2026: payments are no longer just getting faster or cheaper. They’re getting smarter, more interconnected and, fundamentally, more demanding. The five trends below aren’t independent shifts happening in parallel. They’re compounding each other. Miss one, and the others start to feel abstract. Understand how they connect, and you’ll see exactly where the real competitive leverage is.
For the last decade, payments innovation has mostly been about digitisation, moving transactions from paper to screen, from batch to real-time, from one country’s rails to another. That work isn’t done, but it’s no longer the differentiator.
What’s shifting now is the layer above digitisation: how all these systems talk to each other. Card networks talking to AI agents. Stablecoin rails are talking to traditional bank settlement. Authentication systems talking to fraud engines. Domestic instant payment schemes are talking to cross-border corridors. The companies that crack interoperability will be structurally harder to replace.
That’s the through line. Keep it in mind as you read through each trend.
This one moves fast. AI agents, software that shops, compares and transacts on behalf of consumers, are no longer theoretical. OpenAI, Google, Visa and Mastercard all shipped protocols or infrastructure for agentic commerce in the last twelve months. The race is on, and it’s reshaping every assumption payments teams have about how a transaction starts.
The immediate challenge isn’t whether consumers will use these tools. Surveys suggest roughly 81% are open to it, and early data puts the average willingness to spend per agent-enabled purchase around $223 in the US. The challenge is trust and accountability. When a bot buys on someone’s behalf, who verifies intent? Who’s liable if the purchase is wrong? Who owns the fraud risk?
Mastercard answered part of this with Agentic Tokens; tokenisation tied directly to individual users, so agents can transact without exposing payment credentials. OpenAI built a shared protocol (the Agentic Commerce Protocol) that acts as a common language between the AI agent, the merchant and the bank. Google is pushing its own Universal Commerce Protocol. Multiple standards are emerging simultaneously, which creates short-term complexity but signals that the infrastructure is being built seriously.
Stablecoin transaction volumes hit roughly $33 trillion in 2025. That’s more than 20x PayPal’s volume and approaching ACH scale. More importantly, the composition of that volume is changing. A year ago, most of it was trading and arbitrage. Now, genuine payment use cases, such as cross-border B2B settlement, treasury optimisation and contractor payouts, are growing fast enough to matter.
The regulatory clarity that institutions needed finally arrived. The US passed the GENIUS Act, creating a federal framework for payment stablecoins. The EU has MiCA. Banks are no longer watching from the sidelines. JPMorgan, Barclays and others are building stablecoin capabilities or acquiring them.
But here’s what most coverage misses: stablecoins are winning because cross-border payments are still genuinely broken for a lot of use cases. For a company making high-volume payments to suppliers in emerging markets, traditional correspondent banking can take days and cost 5–10x more than domestic transfers. Stablecoins settle in seconds for fractions of a cent.
The friction points that remain are real user experience around on/off-ramps, custody complexity, and lingering trust issues after past depegging events. Smaller businesses are actually leading adoption (around 50% usage among companies with 50–100 employees), while larger enterprises stay cautious. That’ll flip as compliance tooling matures and as more traditional payment players embed stablecoin settlement invisibly into existing flows.
No one writes headlines about data quality, but everyone should be worried about it.
In April 2026, Visa launches its Digital Commerce Authentication Program (VDCAP) in the US and Canada. Merchants who submit specific data fields, Device ID, IP address, email and billing address will qualify for a fee reduction of 0.05%, rising to 0.10% when combined with network tokens. That sounds small until you’re processing at scale.
AI makes this worse before it makes it better. Fraud models are only as accurate as the data they’re trained on. Garbage in, garbage out, except in payments; the output is a declined transaction and a frustrated customer. The teams that invest in clean, complete data flows now will see their approval rates and fraud detection accuracy diverge from competitors who treat data hygiene as an afterthought.
Fraud isn’t growing because attackers got smarter overnight. It’s growing because authentication hasn’t kept pace. SMS One-Time-Passwords (OTPs) were the standard second factor for years. They’re also trivially interceptable via SIM-swap attacks and social engineering. Regulators have noticed.
The UAE banned SMS and email OTPs for financial institutions entirely, with a March 2026 deadline. India follows in April. The Philippines by June. The EU Digital Identity Wallet will roll out by the end of the year. These aren’t just pilot programs; they’re mandates from the government.
Passkeys are the answer the industry has settled on. A payment passkey authenticates a transaction the way you unlock your phone: fingerprint or facial scan, with cryptographic proof staying on the device. No biometric data leaves the hardware. Visa launched Payment Passkeys. Mastercard is retiring the 16-digit card number entirely by 2030, replacing it with tokenisation plus on-device biometrics. Both card networks see this as the authentication layer for the next decade.
The deeper shift is that identity and payment credentials are converging. For years, they were separate systems. Your bank verified you were you, and then your payment network processed the transaction. Now they’re being stitched together. Payment passkeys, digital wallets and mobile driver’s licenses are all moving toward a single portable identity that works across services. For merchants and payment teams, this means authentication friction at checkout is about to drop significantly. But it also means the identity infrastructure you build on needs to be interoperable, not proprietary. Lock-in here will cost you as standards consolidate.
Every trend above depends on this one. Agentic commerce needs protocols that work across AI platforms, merchants and banks simultaneously. Stablecoins need on/off-ramps that connect to traditional payment rails. Data quality mandates need standardised messaging (ISO 20022) to flow consistently across borders. Digital identity needs portable credentials that work everywhere.
Interoperability is unsexy. It’s also the single biggest determinant of who wins the next phase of payments.
For companies operating globally, the implication is stark. Fragmented payment stacks, one provider for Europe, another for Asia, a third for the Americas, are becoming a liability.
The trend is toward consolidation: single platforms that connect bank accounts, wallets, stablecoin endpoints, and local payment rails under one API. Nomupay, is building exactly this. The companies that can serve as genuine bridges between traditional and emerging rails will own the next wave of cross-border commerce.
Pull your current data submission into compliance with Visa’s VDCAP requirements before April. See exactly which fields you’re missing and what that’s costing you in approval rate today, not when the mandate hits.
Spend one hour this week reading the OpenAI Agentic Commerce Protocol documentation. You don’t need to build anything yet. But you need to understand the architecture before your competitors do.
Learn more about being ready via a payment processor at Nomupay.
