Ant International expands AI-commerce payments infrastructure

According to The Manila Times and the South China Morning Post (SCMP), Ant International now connects over 150 million merchants with more than 2 billion consumer accounts globally, and the network supports 300+ payment methods across 220 markets while processing over 20 million transactions daily on average (The Manila Times; SCMP). The Manila Times lists multiple national QR systems and recent partner onboardings in Asia, Latin America and the Middle East, and notes collaborations with local networks such as Vietnam's NAPAS and Mexico's R2 (The Manila Times). SCMP reports that at MoMents 2026 the company unveiled an open-sourced Agentic Mobile Protocol and announced a business-facing AI payment processing product via Alipay (SCMP). SCMP also reports that Alipay's consumer-facing AI Pay launched in September and, per the outlet, surpassed 100 million users and processed more than 120 million transactions in one week in February (SCMP).

Industry-pattern observations: open-sourcing a payments-layer protocol for agent interactions reflects a broader trend toward composable, machine-consumable interfaces in fintech. Comparable initiatives in payments and identity aim to replace human-driven web flows with API- and agent-native exchanges, which typically raise integration, security, and interoperability trade-offs that platform operators and integrators handle via standardized protocols and SDKs.

Editorial analysis: the combination of large-scale payment rails and agent-facing protocols could lower friction for automated, pay-per-use commerce workflows used by AI assistants, marketplaces, and embedded finance providers. Observers note that merchant onboarding, settlement timing, fraud detection, and reconciliation are common operational bottlenecks when moving from manual to agent-driven payments, based on prior rollouts in embedded and cross-border payments.

Editorial analysis: practitioners and integrators should monitor:

• •adoption and implementation details of the Agentic Mobile Protocol
• •how tokenization, consent, and authentication are handled for agent-initiated transactions
• •settlement and reconciliation latency for pay-per-use models. Public testing, reference implementations, and developer tooling will be leading indicators of practical utility