Bitcoin $82,212.00 +0.85%

Ethereum $2,438.56 +0.88%

Tether $0.9991 −0.01%

XRP $1.33 +0.65%

Solana $105.30 +1.06%

Cardano $0.2302 +0.48%

Dogecoin $0.0811 −0.58%

Litecoin $61.13 +3.11%

Chainlink $12.46 +0.23%

Polkadot $1.15 +0.30%

Blockchain Data Wiki Crypto Blog
Payments

AI Agents Drive 138 Million Payments on Coinbase's x402 Protocol

  • 3
  • 3 min read

A new report reveals that AI agents have processed over 138 million payments through the x402 protocol, with the US leading the charge.

Recent findings from Chainalysis have unveiled a significant leap in the utilization of cryptocurrency by artificial intelligence (AI) agents. These autonomous programs have completed more than 138 million transactions via the x402 protocol hosted on Coinbase's Base blockchain. This evolution marks a new era where AI can facilitate payments for digital resources without any human oversight.

The Rise of x402 Payments

The x402 protocol, nurtured by Coinbase, allows for automated payments using cryptocurrency, enabling AI agents to procure services such as data access and computing power. Unlike conventional payment methods, x402 can handle ultra-small transactions, thus eliminating high processing fees that typically apply to traditional networks.

Chainalysis' analysis indicates that a staggering 87% of payments made via x402 were small transfers, typically ranging from just $0.0001 to $0.10, amounts that standard credit card systems would struggle to process affordably.

Stablecoins as a Key Resource

Stablecoins are emerging as an essential medium for AI agents to purchase digital services. As highlighted by Stephanie Cohen, Cloudflare's chief strategy officer, these tokens are particularly well-suited for the kinds of transactions that autonomous systems are starting to make.

Major players in the financial services industry, including Visa, Mastercard, and Stripe, are recognizing the potential of the x402 protocol, with Coinbase facilitating business customers in accepting payments from AI agents.

The US vs. China: A Payment Landscape Divide

There is a clear geographic disparity in the payments for AI inference services. Notably, US-linked wallets comprised a remarkable 88.1% of identified transactions, while those associated with China represented a mere 0.1%. This trend suggests divergent usage patterns in both regions.

Eric Jardine, Chainalysis' head of research, notes that this difference may stem from Chinese developers relying more on open-source AI models, which could lead to a lower demand for paid AI services. Despite this, he advises caution in interpreting these statistics as a decline of on-chain inference markets in Asia.

However, it's worth mentioning that Chinese-linked wallets excelled in payments for security and code-analysis services, accounting for about 90% of that activity, predominantly through services like AgentLisa, which is designed for smart contract audits.

Trends in Wallet Activity

In addition to examining payment volumes, Chainalysis has tracked the speed of transactions from newly funded wallets. For instance, among wallets that received funding in the second quarter of 2026, 89.3% continued to receive further funding within just 30 days. This contrasts with a lower rate of 70.6% for wallets funded in the third quarter of 2025. Moreover, the average time from initial funding to the first x402 payment significantly dropped from 73 days in late 2025 to a mere 18 minutes by the third quarter of 2026.

Industry experts like Yat Siu from Animoca Brands predict that the number of AI agents conducting transactions online could reach up to 100 billion in the coming years, further emphasizing the growing reliance on cryptocurrencies for these transactions.

Implications and Conclusion

The Chainalysis report underscores the dynamic nature of AI transactions, revealing over 138 million x402 payments to date, primarily driven by US-linked wallets. While China has dominated specific areas like security services payments, faster wallet funding and a prevalence of micropayments are indicative of evolving transaction habits.

While these figures show an increase in activity, they don't entirely capture the demand dynamics or distinguish between genuine commercial use and automated transactions devoid of economic engagement.