When AI learns to pay for itself, the big companies are already quietly competing.

Source: Tiger Research

Authors: Ekko, Ryan Yoon

Original Title: AI Agent Payment Infrastructure: The Direction of Crypto and Big Tech

Translation and Editing: BitpushNews

An era driven by AI and led by automation is approaching. To make automation truly “autonomous,” it must have native payment capabilities. The market has already begun laying the groundwork for this transition.

Key Points

• The payment subject is shifting from humans to AI Agents, making payment infrastructure a core requirement for achieving true autonomy.
• Tech giants (including Google AP2 and OpenAI’s delegated payments) are designing approval-based automated payment systems on top of existing platform infrastructure.
• Cryptocurrencies (via ERC-8004 and x402) leverage NFT-based identity recognition and smart contracts to enable trustless payment models.
• Tech giants prioritize convenience and consumer protection, while cryptocurrencies emphasize user sovereignty and broader agent-level execution capabilities.
• The key question for the future is: Will payments be controlled by platforms or executed via open protocols?

1. Payments Are No Longer Exclusive to Humans

Source: macstories (by Federico Viticci)

Recently, “OpenClaw” has garnered widespread attention. Unlike major AI systems like ChatGPT or Gemini, which focus on retrieving and organizing information, OpenClaw enables AI Agents to execute tasks directly on the user’s local PC or server.

Through messaging platforms like WhatsApp, Telegram, and Slack, users can issue commands, and the Agent autonomously performs tasks including email management, calendar coordination, and web browsing.

Because it runs as open-source software and is not tied to a specific platform, OpenClaw functions more like a personal AI assistant. Its architecture is favored for its flexibility and user-level control.

However, a key limitation remains: for AI Agents to be fully autonomous, they must be able to execute payments. Currently, Agents can search for products, compare options, and add items to carts, but final payment authorization still requires human approval.

Historically, payment systems were designed around human actors. In an AI-driven environment, this assumption no longer holds. For automation to be fully autonomous, Agents must be able to evaluate, authorize, and complete transactions independently within defined constraints.

Anticipating this shift, major tech giants and native crypto projects have launched technical frameworks over the past year aimed at enabling agent-level payments.

2. Big Tech: Building Agent Payments on Existing Infrastructure

In January 2025, Google launched AP2 (Agent Payment Protocol 2.0), expanding its AI Agent payment infrastructure. While OpenAI and Amazon have also outlined related plans, Google is currently the only large company with a structured implementation framework.

AP2 divides the transaction process into three authorization layers (Mandate Layers). This structure allows independent monitoring and auditing of each stage:

1. Intent Mandate: Records what the user wants to do.
2. Cart Mandate: Defines how to execute the purchase based on preset rules.
3. Payment Mandate: Executes the actual transfer of funds.

Example Scenario: How Google AP2 Works

Suppose Ekko requests an AI Agent on Google Shopping to “find and buy a winter jacket under $200.”

• Intent Mandate: Ekko instructs the AI Agent to purchase “a winter jacket with a maximum budget of $200.” This information is recorded as a digital contract on-chain, called the intent mandate.
• Cart Mandate: The AI Agent follows the intent, searches partner merchants for products matching “winter jacket” and “under $200,” and adds qualifying items to the cart.

“Selected item: winter jacket,” “Price verified: $199 (within budget ✓).”

“Added to cart,” “Shipping address confirmed.”

• Payment Mandate: Ekko confirms the AI Agent’s selected item and clicks the payment approval button. The $199 is processed via Google Pay. Alternatively, the AI Agent can automatically complete the payment within predefined parameters.

Throughout this process, the user does not need to input additional information. Google AP2 operates on top of Google Pay, utilizing pre-registered card details and shipping addresses. Because AP2 relies on existing user credentials, it reduces onboarding friction and simplifies adoption.

Source: Google

However, Google currently only supports agent-based payments within its partner network. Its scope remains limited to a controlled ecosystem, restricting broader interoperability and open access.

3. Cryptocurrency: Self-Hosting and Open Exchange

The crypto space is also developing payment infrastructure for AI Agents, but with a different approach. Large platforms build trust within controlled ecosystems, whereas crypto starts from a different question: can AI Agents be trusted without relying on centralized platforms?

Two core standards aim to achieve this: Ethereum’s ERC-8004 and Coinbase’s x402.

Identity and Payment Integration

First, consider the identity layer. Just as humans need IDs to access digital services, AI Agents operating on blockchain networks must be recognizable. ERC-8004 serves this purpose.

It issues NFTs—not as media collectibles, but as credentials containing structured identity data. Each token includes three components:

1. Identity
2. Reputation
3. Validation

These elements form a verifiable on-chain identity certificate. In e-commerce, participants review ratings and transaction history before trading; the same logic applies to AI Agents. ERC-8004 provides verifiable credentials for Agents, enabling others to assess whether a transaction is appropriate based on transparent data.

However, identity alone does not enable value transfer; a payment mechanism is also needed. This role is fulfilled by x402.

If ERC-8004 is a digital ID, then x402 is the payment channel. Developed by Coinbase, x402 is a crypto-native payment standard for AI Agents. It allows Agents to autonomously transact using stablecoins.

Its core function is automating smart contract execution. Conditions like “auto-transfer upon meeting predefined criteria” are embedded directly in code. Once conditions are met, settlement occurs without human intervention.

When combined, ERC-8004 for identity and x402 for payments enable AI Agents to verify counterparties and execute transactions trustlessly. Trust and settlement are handled at the protocol level, not controlled by platforms.

Example Scenario: Business Between Agents Using ERC-8004 and x402

Imagine a near-future AI environment: Ekko instructs his AI Agent (Agent A) to purchase a used laptop with a maximum budget of $800. The market operates its own AI Agent (Agent B), which communicates directly with Ekko’s Agent to execute the transaction.

1. Mutual Verification:

Before trading, both Agents verify each other’s credentials and confirm product compliance.

• Identity check: verified via ERC-8004 NFT
• Ekko’s Agent: reputation score 72, confirmed balance $800
• Seller’s Agent: reputation score 70, confirmed inventory of qualifying laptops
• Result: Both Agents are authorized to proceed.

2. Escrow:

After verification, the transaction begins. Each Agent interacts via x402 protocol to transfer and confirm funds.

• Escrow: $800 transferred from Ekko’s Agent wallet to a smart contract.
• Conditional Lock: Funds remain locked until delivery is confirmed.
• Release: After delivery confirmation, $800 automatically transfers to the seller.

3. Settlement and Reputation Update (via x402 settlement and NFT update):

Post-settlement, both Agents’ reputation records are updated.

• Ekko’s Agent: reputation 72 → 80 (+5 for quick delivery, +3 for accurate description)
• Seller’s Agent: reputation 70 → 78 (+5 for quick delivery, +3 for accurate description)
• The updated ratings are recorded in each Agent’s ERC-8004 NFT.

Throughout, no intermediaries are involved, and no platform approval is needed. The two AI Agents transact directly through blockchain-based verification and settlement, exemplifying a crypto-native agent-to-agent commerce model.

4. Big Tech vs. Cryptocurrency: Differences in AI Agent Operating Domains

Control vs. Openness

Google AP2 exemplifies a controlled model designed for approved partners.

Google restricts market participation to vetted merchants to protect consumers. Even with a structured authorization framework, Agent behavior cannot be fully guaranteed. Unlike deterministic systems with direct input-output matching, AI Agents produce probabilistic results.

If an Agent connects to an unreliable partner and a transaction error occurs, liability may ultimately fall on the payment infrastructure provider. To reduce failure probability even by 0.01%, Google has incentives to narrow its ecosystem. This controlled environment enhances stability and regulatory oversight but may limit Agents’ ability to operate autonomously across broader markets and optimize across multiple options.

In contrast, ERC-8004 and x402 reflect a more open architecture. The crypto approach aims for permissionless interoperability, not platform dependency.

Efficiency and Use Cases

AI Agents are still in early development. Seamless end-to-end execution—from complex requests to autonomous payments—is not yet fully realized. However, the long-term vision is for Agents to independently manage routine consumption. For example, a user might instruct an Agent to restock groceries, which then assesses inventory gaps and automatically completes purchases.

Major platforms may attempt to aggregate key retail channels to support such models within a unified environment. This approach can enable reliable daily use within a controlled framework. But integrating all potential counterparties—including small online merchants, independent websites, DeFi protocols, and exchanges—faces structural limitations in closed ecosystems.

Furthermore, as digital content increasingly shifts to paid access, Agents may need to perform high-frequency microtransactions. Open crypto standards could have structural advantages here. For instance, an AI Agent could buy 1,000 creator-generated images at $0.01 each or pay $1 to access a research paper. For small, programmable payments, crypto-native protocols may offer higher operational efficiency.

That said, the absence of centralized institutions also introduces trade-offs. Identity verification standards must be established in a decentralized manner, with no single entity bearing ultimate responsibility for failures. Balancing openness with accountability remains a key design challenge, contingent on technological maturity and usability improvements.

Summary

Big tech giants and the crypto space are both pursuing the same goal: enabling autonomous AI Agent commerce. The difference lies in architecture. Large corporations tend toward closed, controlled systems, while crypto advocates for open, protocol-based models.

This is unlikely a zero-sum game; a more probable trajectory involves interoperability between these approaches. In the current stage of technological development, ongoing efforts should prioritize reliability and user experience.