The July 2026 Wave of Announcements
On 20 July, the Emerging Payments Association Asia launched an AI and Agentic Payments Working Group with HSBC as a founding member, aiming to define standards for agentic commerce across Asia Pacific. Days earlier, the Linux Foundation established the x402 Foundation to develop an open, vendor-neutral payment standard built over HTTP, designed to make internet-native payments legible to AI agents. UnionPay, meanwhile, unveiled an Agentic Payment Open Protocol at the World Artificial Intelligence Conference in Shanghai, and finance-focused vendors such as Payouts.com rolled out role-based AI agents for treasury and accounts-payable teams.These followed Mastercard’s earlier launch of Agent Pay, a protocol allowing AI agents to make payments and send micropayments to one another. Taken together, the announcements point to a field consolidating quickly around a shared question: who is liable, and how is trust established, when an agent transacts on a user’s behalf?How Agentic Payments Standards Are Forming
The core technical challenge is identity and authorisation. A human checkout relies on a person present to approve a charge; an agentic payment must encode consent, spending limits and accountability into the transaction itself. Competing efforts are approaching this from different angles: open web standards like x402, card-network protocols from Mastercard and UnionPay, and bank-led working groups such as the EPAA-HSBC initiative focused on liability. Where crypto rails enter the picture, stablecoins are an obvious settlement layer for machine-to-machine payments because they clear instantly and around the clock, a role reinforced by regulatory milestones such as Circle’s national trust bank approval for USDC.Why Institutions Are Moving Now
Financial institutions are engaging early for a defensive reason as much as an offensive one: whoever sets the standard for agentic payments captures the routing, data and trust layer for a large share of future commerce. HSBC and Mastercard have already piloted end-to-end B2B agentic transactions for Singapore-based clients, and the International Monetary Fund has published analysis on how agentic AI could reshape payments. The involvement of incumbents, rather than only startups, is what distinguishes this cycle from earlier automation hype.What It Means
The near-term reality is a mix of live pilots and competing draft standards rather than a single settled system. For institutions, the practical takeaway is to track which protocols gain traction and to treat identity, consent and liability as the decisive design questions. For the crypto sector specifically, agentic payments strengthen the case for stablecoins and programmable settlement as the rails machines will prefer. The risk is fragmentation: multiple overlapping standards could slow adoption, and the liability frameworks are still unwritten. As with the broader market, readers should weigh this against a still-volatile backdrop, covered in our Bitcoin price outlook for the second half of 2026.Frequently Asked Questions
What are agentic payments? Agentic payments are transactions initiated, authorised and settled by autonomous AI agents acting on a user’s behalf, rather than by a person manually approving each charge.Which organisations are setting agentic payments standards? Efforts include the Linux Foundation’s x402 Foundation, Mastercard’s Agent Pay, UnionPay’s Agentic Payment Open Protocol, and the EPAA working group co-founded with HSBC.📧 The Gulf reads Cryptonite first
Get MENA regulation moves, RWA deals and AI-money trends in one weekly brief — plus instant alerts when the MENA Regulation Tracker changes. Free, no spam.
Was this briefing useful?Thanks for the feedback!
