Du Xiaoman Launches ClawPay to Compete in the AI Agent Payment Race

As AI Agents evolve from "chatting" to "executing," payment capabilities are emerging as a critical new battleground for global Technology companies and financial institutions.

On May 13, during Baidu Create 2026, Du Xiaoman unveiled ClawPay, a payment solution designed for AI Skill developers. The product encapsulates billing, order placement, payment initiation, and account management into Standardized technical services, enabling developers to quickly integrate monetization capabilities into their AI Skills without building their own payment infrastructure.

With open-source Agent ecosystems like openclaw rapidly gaining trACTion, the global AI payment sector has entered a period of intense competition over protocols and standards. Internationally, giants such as Google, Stripe, Coinbase, and Visa have launched AI payment protocols, while domestically, Ant Group, WeChat Pay, JD Technology, and China UnionPay are making significant moves.

In this industrial shift from "human-led payments" to "machine-agent payments," security, compliance, and a trustworthy architecture will be the core competitive moats.

ClawPay: Completing the AI agent Transaction Loop

Based on publicly available Information, ClawPay's core goal is not merely to "help developers collect money," but to address a critical bottleneck in the current agent economy: how AI APPlications complete the loop from capability dEMOnstration to real-world transactions.

Over the past year, the global AI Agent ecosystem has ballooned. Autonomous agent platforms like OpenClaw have gained popularity, with many agents now capable of invoking tools, accessing local files, executing tasks, and even autonomously operating internet services. OpenClaw has become one of the most watched open-source agent projects of 2026, with its ecosystem beginning to integrate high-permission services such as email, payments, and file systems.

However, as AI agents enter the commercial world, a practical problem has quickly surfaced: agents can complete tasks but struggle to finalize "transactions." For many AI Development teams, the complexity of payment integration has long been the most difficult part of commercialization. Developers not only need to write custom logic for order placement, billing engines, and payment interfaces but also adapt to the distinct technical specifications and compliance requirements of various channels like WeChat Pay and Alipay.

ClawPay aims to platformize this capability. According to Du Xiaoman, ClawPay supports major payment methods, including WeChat Pay, Alipay, and Du Xiaoman Pay, and unifies order management, billing engines, and payment components. Developers simply configure information like unit prices and subscription packages, and the platform automatically generates the complete monetization logic, listing it on the OpenClaw Skill Store. The process can be completed in as fast as 10 minutes, featuring "zero code, zero setup fees, and zero platform service fees."

International Giants Vie for Dominance

International tech and payment behemoths are already locked in fierce competition globally. google's offensive has been particularly aggressive. In the fall of 2025, Google partnered with over 60 institutions, including PayPal, Stripe, Ant International, Coinbase, and Mastercard, to launch the open-source Agent Payments Protocol (AP2), which introduces tamper-proof cryptographic "Letters of Authorization" and is compatible with both traditional credit cards and stablecoins. In January 2026, Google released the Universal Commerce Protocol (UCP), aiming squarely at DeFining the commerce protocol layer for the AI agent economy. Its ambition extends beyond payments to becoming the standard-setter for the entire agent-driven commercial ecosystem.

Payment company Stripe collaborated with OpenAI on the Agentic Commerce Protocol (ACP), allowing ChatGPT users to complete shopping and payments directly within the conversational interface. In March 2026, Stripe joined forces with blockchain firm Tempo to launch the Machine Payments Protocol (MPP), supporting Micropayments as low as 0.01 USDC, catering to the high-frequency, small-value nature of AI transactions. The crypto payment sector also views the Agent economy as a strategic market. In May 2025, Coinbase initiated the x402 Protocol with Cloudflare, Google, and Visa, activating the HTTP 402 standard as an on-chain payment layer for AI agents. In February 2026, it launched Agentic Wallets, enabling AI agents to hold funds independently and execute transactions autonomously, marking agent payments as a top corporate priority. Stablecoin giant Circle launched a USDC smart contract layer, exploring the use of stablecoins for micropayments in AI agent scenarios, with CEO Jeremy Allaire stating in 2025 that the combination of AI and stablecoins would be pivotal in the machine Economy. Card networks like Visa and Mastercard have also been advancing research on "Agentic Commerce," envisioning future AI assistants making certain consumption decisions on behalf of users.

Intensive Strategic Moves Domestically

China's financial system is simultaneously laying the groundwork. In July 2025, Ant Group took the lead by opening its MCP interface, allowing third-party agents to access Alipay's payment capabilities and connecting the entire chain of "agent invocation, service pricing, user confirmation, automatic payment." By February 2026, Alipay's "AI Pay" user base had surpassed 100 million, with cumulative weekly transactions exceeding 120 million, making it the world's first AI-native payment product to surpass the 100-million threshold in both metrics. Around the Same time, WeChat Pay integrated with Tencent Yuanqi, leverAGIng its super-ecosystem of over 1.3 billion monthly active users to offer unmatched user reach in AI payment scenarios.

Between late March and early April 2026, Ant Group and WeChat Pay successively launched "Payment Skills," signaling an escalation of the AI payment standards war from the protocol level to the application level. On March 31, 2026, JD Technology introduced ClawTip, an A2A micropayment infrastructure based on the x402 protocol, granting AI "decision-making payment" capability—allowing an agent to autonomously judge, inquire, and complete a payment within a pre-set sandbox limit, rather than simply executing a pre-authorized transaction. ClawTip adopts a zero-fee model with completely free payment, collection, and withdrawal services, attempting to woo the developer ecosystem by lowering monetization bARRiers. On April 2, 2026, China UnionPay officially launched its Agent Payment Open Protocol (APOP) Framework in Shanghai, completing five live production verification transactions. In the demonstration, a user purchased flight tickets via voice command within a travel AI Assistant, which automatically linked user information and debited a Bank of Communications credit card. The APOP framework encompasses four core capabilities: agent identity management, intent management, user identity management, and payment authorization management, attracting 19 domestic and international institutions, including commercial banks, agent manufacturers, and acquirers.

The Scramble for a New gateway

These developments signal the payment industry's transition from "Internet payments" to "machine-agent payments." In traditional online payments, the principal is always a human; in the agent era, a user may simply express an intent, with subsequent price Comparisons, ordering, payment, and after-sales service handled automatically by the AI. Payment systems are thus evolving from simple clearing utilities into "permission-control systems" within AI architectures. Consequently, security and compliance are becoming the core competitive differentiators for the industry. Du Xiaoman emphasized that ClawPay retains familiar authentication habits like passwords and biometric verification and generates a full-chain audit snapshot for every transaction record to meet traceability and auditing requirements. This focus is no accident. Multiple recent studies have identified significant security risks in high-permission agents. Research on OpenClaw reveals that agents with access to payments, email, and file systems are susceptible to prompt injection attacks, privilege escalation, and identity spoofing. Some studies even suggest that the vulnerabilities in current agent systems are "architectural-level risks." As a result, the race to be the first to build a "trusted agent payment system" has become the new gateway that global tech companies, payment providers, and financial networks are all vying to control.