XAgent is the open execution market for Agentic Commerce. It turns merchant products, pricing, inventory, order logic, settlement, and fulfillment capability into commerce interfaces that AI agents can read, call, and purchase from. Agents quote, authorize, settle, fulfill, and prove the purchase — machine-to-machine.

What is agentic commerce?

Agentic commerce is a model where AI agents — not humans clicking through web pages — discover services, place orders, and settle payments autonomously. XAgent makes this real through five execution pillars: Quotable, Authorizable, Settleable, Fulfillable, Provable.

How is XAgent different from Stripe or Coinbase Commerce?

Stripe was built for humans clicking checkout buttons. Coinbase Commerce is a crypto payment gateway. XAgent is the protocol-native execution market where AI agents transact: protocols standardize the request, XAgent operationalizes the transaction. The moat is merchant operations — catalog, quote, authorization, order, settlement, delivery, receipt, refund, reconciliation, fraud signal, jurisdiction metadata.

How do AI agents pay merchants on XAgent?

Agents pay using the x402 protocol — an HTTP 402 challenge/response standard. The agent receives a signed payment instruction, authorizes a stablecoin transfer (USDG or USDT) via EIP-3009, and the merchant is paid in seconds with zero gas fee to the buyer (OKX facilitator covers it).

Do I need to write code to list my store?

No. One-click OAuth in your Shopify or WooCommerce admin turns your store into a merchant capability that AI agents can quote, order, pay, settle against, and prove — zero code, zero contracts.

Agent payment went live on three continents in one week

Alipay, Robinhood, Crossmint, and Worldline shipped agent payment products between May 26 and June 2. The convergence is the signal. Merchant operations is the bottleneck.

By XAgent Team · 2026-06-05

Between May 26 and June 2, 2026, six companies on three continents shipped agent payment products into production. Alipay launched an AI Wallet in China. Robinhood gave agents a trading account and a virtual credit card in the United States. Crossmint shipped an Agentic Cards API on Visa rails. Worldline, ING, and Mastercard completed Europe's first live end-to-end agentic payment. The products are different. The conclusion is the same: agentic commerce is now a global infrastructure race, and the bottleneck is shifting from "how does an agent hold money" to "how does a merchant handle an agent buyer."

Seven days of agent payment, from Hangzhou to Amsterdam

On May 26, Alipay unveiled its full-stack AI payment infrastructure at a dedicated event in Hangzhou. The AI Wallet gives users control over agent spending before, during, and after a task. Token Pay handles subscription billing and token top-up for AI model providers. Alipay disclosed that its AI Pay product had already crossed 100 million users and processed 300 million agent-initiated transactions — numbers no other agent payment product has reached.

One day later, Robinhood announced two parallel products: agentic trading with a separate agent account and dedicated wallet balance, and an agentic virtual credit card for Gold Card holders. Users set spending limits and choose whether each transaction requires approval.

The same day, Highnote integrated Visa Intelligent Commerce into its card-issuing SaaS platform, targeting B2B workflows — invoice automation, vendor payments, procurement — where agents operate inside corporate spend policies.

On June 2, Crossmint launched its Agentic Cards API using Visa Intelligent Commerce and Basis Theory. Developers can provision tokenized Visa credentials for agents, with card numbers and CVCs held outside the agent runtime. The product is already live in lobster.cash, a tool that installs into Claude Code and other agent platforms.

The same day in Amsterdam, Worldline, ING, and Mastercard announced Europe's first live, production-environment, end-to-end agentic payment. An ING cardholder's agent found concert tickets within budget, presented filtered options, and completed the purchase after explicit consumer approval. The transaction carried an identifier marking it as agentic, giving ING full visibility in the authorization flow.

Six products. Seven days. Three continents.

Every product converged on the same agent payment model

These companies did not coordinate. They built in different markets, for different users, on different rails. Yet they arrived at the same four-part architecture:

Isolated account. The agent operates from a separate balance, wallet, or credential — not the user's full account. Alipay uses a dedicated AI Wallet. Robinhood creates a distinct agent account. Crossmint provisions a scoped token.
Bounded authorization. The agent's spending capability is constrained by amount, category, merchant, or task type. Every product ships with configurable limits.
Conditional approval. Some transactions run autonomously within preset rules. Others require per-transaction human confirmation. Every product supports both modes.
Post-transaction audit. Each transaction produces a record the user can review — spending summaries, order IDs, settlement proofs.

This is convergence on the minimum viable trust model for agent payment. Experian's research supports the demand side: 55% of consumers said they would allow an AI agent to make an autonomous purchase on their behalf, rising to 70% among 25- to 39-year-olds. But willingness requires guardrails — and every builder independently concluded that the guardrails look the same.

The convergence signals one thing: the credential layer is getting solved. Six companies now ship scoped agent wallets. More will follow. That means more agents will attempt to buy things.

The question is whether merchants are ready to sell to them.

Agentic commerce needs merchant operations, not just credentials

A credential answers one question: can this agent attempt to pay? A real purchase needs several more answers before that payment should happen.

When Worldline's agent found concert tickets for the ING cardholder, that worked because a specific merchant had been pre-integrated into the pilot. The agent knew which product catalog to query, which prices were current, which payment method to use, and how to present options for consumer approval. Worldline CEO Madalena Cascais Tomé called it "production-ready" — and for one merchant in one flow, it was.

Scale that to a thousand merchants, and the integration problem becomes an operations problem:

Which catalog does the agent read? Is it the same data the merchant's website shows, or a separate feed? Does it include inventory, variants, shipping, and regional availability?
When does a price become a binding quote? Is the quote time-bounded? Does it include taxes, fees, and currency conversion?
Which agent mandates does the merchant accept? What happens when an agent's spending limit falls between two product tiers?
How does settlement connect to order creation? Does the merchant's order system know the payment came from an agent?
What proof does the agent return to the user? A transaction hash? An order ID? A delivery tracking number?

These are not payment questions. They are merchant operations questions. No credential — no matter how well-scoped — answers them.

This is the layer XAgent is built around. We describe XAgent as the open execution market for agentic commerce because the market is not a list of merchants. It is the operating surface that makes each merchant discoverable, quotable, authorizable, settleable, fulfillable, and provable. A merchant connects once and becomes agent-executable across protocols and payment rails. The agent does not need to know whether the merchant prefers x402, MCP, or a card network flow. The execution layer handles the difference.

Protocol-native, not protocol-limited.

What's next

Agent credentials will keep shipping. Visa and Mastercard are both expanding their agentic programs. Crossmint is working with Mastercard and American Express to extend agentic cards beyond Visa. Robinhood is adding options, crypto, and futures to its agentic trading. Alipay's Token Pay is onboarding more AI model providers.

The more credentials ship, the more agents will try to buy. And the more agents try to buy, the more merchants will discover that accepting a card payment from a human is not the same as accepting a purchase from an agent.

We are building the merchant operations layer that closes that gap. If your business needs to be bought by agents — not just browsed by them — list your store and make the transaction executable.