Tiger Research: How Crypto Giants Are Betting on AI Agent Payment Infrastructure

This report is authored by Tiger Research. To achieve true autonomous automation, native payment capabilities are essential. The market has already begun actively preparing for this shift.

Key Points

The payment主体 is shifting from humans to AI Agents, making payment infrastructure a core requirement for true autonomy.

Major tech companies (including Google AP2 and OpenAI Delegated Payment) are designing approval-based automated payment systems on top of existing platform infrastructure.

Cryptocurrencies, through standards like ERC-8004 and x402, leverage NFT-based identity verification and smart contracts to realize decentralized payment models.

While big tech prioritizes convenience and consumer protection, 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. Payment is no longer exclusive to humans

Source: macstories (Federico Viticci)

Recently, “OpenClaw” has garnered widespread attention. Unlike AI systems like ChatGPT or Gemini, which primarily retrieve and organize information, OpenClaw allows AI Agents to execute tasks directly on users’ local PCs or servers.

Through messaging platforms like WhatsApp, Telegram, and Slack, users can issue commands, and the Agent autonomously performs tasks such as 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 critical 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主体. In an environment driven by AI Agents, this assumption no longer holds. For automation to be truly autonomous, Agents must be able to independently evaluate, authorize, and complete transactions within defined constraints.

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

2. Major tech companies: 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 initiatives, Google is currently the only large enterprise with a structured implementation framework.

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

Intent Mandate: Records the user’s desired action.

Cart Mandate: Defines how to execute the purchase under preset rules.

Payment Mandate: Executes the actual transfer of funds.

Example: Suppose Ekko instructs a Google Shopping AI Agent to “find and buy a winter jacket under $200.”

Intent Mandate: Ekko directs the AI to purchase “a winter jacket with a maximum budget of $200.” This instruction is recorded on-chain as a digital contract, i.e., the intent mandate.

Cart Mandate: The AI follows the intent, searches among partner merchants for matching items, adds suitable products to the cart, verifies the price ($199, within budget ✓), and confirms the delivery address.

Payment Mandate: Ekko reviews the selected item and approves it. The $199 is processed via Google Pay. Alternatively, the AI can automatically complete the payment within preset parameters.

Throughout this process, the user does not need to input additional information. Google AP2 relies on existing user credentials (pre-registered cards and addresses), reducing entry barriers and simplifying adoption.

Source: Google

However, Google currently only supports agent payments within its partner network. This limits its scope to a controlled ecosystem, restricting broader interoperability and open access.

3. Cryptocurrency: Self-custody and open exchange

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

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

First is the identity layer. AI Agents operating on the blockchain must be recognizable. ERC-8004 serves this purpose. It issues NFTs that are not art collectibles but credential tokens containing structured identity data. Each token includes three components:

Identity

Reputation

Validation

These elements together form a verifiable on-chain identity certificate.

For payment mechanisms, x402 functions as the payment pathway. Developed by Coinbase, x402 is a native crypto payment standard for AI Agents. It enables autonomous transactions using stablecoins. Its core feature is automated smart contract execution, with embedded conditional logic—once conditions are met, settlement occurs without human intervention.

When ERC-8004 (identity) and x402 (payment) are combined, AI Agents can verify counterparties and execute transactions without relying on centralized platforms.

Example: Ekko instructs his Agent A to purchase a used laptop with a maximum budget of $800. Agent B, the seller’s agent, communicates directly.

Mutual verification: Using ERC-8004 NFTs, verify identity and reputation scores (e.g., reputation 72, balance confirmed).

Escrow smart contract: $800 is transferred from the wallet into escrow, locking funds until receipt is confirmed.

Settlement and reputation update: After the transaction, x402 automatically settles, and both parties’ reputation records are updated and written into their respective ERC-8004 NFTs.

Throughout this process, no intermediaries are involved. The two AI Agents conduct direct transactions via blockchain-based verification and settlement, exemplifying a crypto-native, agent-to-agent (A2A) business model.

4. Major tech vs. cryptocurrency: Differences in AI Agent operation domains

Google AP2 exemplifies a controlled model designed for verified partners. Google restricts market participation to protect consumers. Because AI Agent execution results are probabilistic rather than fully deterministic, if a transaction error occurs, liability may ultimately fall on the payment infrastructure provider. To reduce failure risk, Google has an incentive to narrow its ecosystem.

This restricted ecosystem enhances stability but also limits the ability of Agents to operate autonomously and optimize choices in broader markets.

In contrast, standards like ERC-8004 and x402 reflect a more open architecture. The crypto model aims for permissionless and interoperable systems.

While end-to-end execution is not yet perfect, the long-term vision is for Agents to independently manage daily consumption. Major platforms might attempt to integrate key retail channels, but open crypto standards have structural advantages in handling small, high-frequency programmable payments (microtransactions). For example, an Agent purchasing 1000 images at $0.01 each via crypto-native pathways can operate more efficiently.

Of course, the absence of centralized institutions also entails trade-offs: identity assessment standards must be established in a decentralized manner, and no single entity bears ultimate responsibility for failures.

Summary

Both big tech and crypto are pursuing the same goal: enabling autonomous AI Agent commerce. The difference lies in architecture: big tech favors closed, controlled systems, while crypto advocates open, protocol-based models.

The future is more likely to involve interoperability between these approaches rather than a zero-sum competition.