AI Agents Transacting on-Chain: Rising Use of x402 and Stablecoins

Autonomous AI agents are increasingly acting as independent economic participants, paying for services, buying datasets and acquiring access tokens for AI models. Chainalysis’ recent report, Inside North America’s $3 trillion crypto economy: Agentic commerce accelerates, documents more than 138 million transactions routed through the x402 protocol on Coinbase’s Base blockchain.

These payments are frequently tiny, often far below what traditional card networks can handle economically, yet they enable machine-to-machine commerce at scale.

U.S. Leads the Emerging Market

The United States dominates this nascent market: wallets linked to the U.S. accounted for 88.1% of geographically identified payments for AI model inference. By contrast, Chinese-linked wallets made up only 0.1% of those payments. Chainalysis suggests one reason is that many Chinese developers favor open-weight models such as DeepSeek and Qwen, which reduce the need to buy metered inference services.

Payments giants including Visa, Mastercard and Stripe have supported the x402 protocol, and Coinbase now lets business customers accept agent-driven payments. Industry voices anticipate a surge in AI agents transacting online — possibly billions — with crypto and stablecoins playing a central role.

Why Stablecoins and x402 Matter

Stablecoins fit agent commerce because they permit transactions worth fractions of a cent, enabling agents to acquire small amounts of compute, data or other microservices without human intervention. Chainalysis’ data show that over 87% of x402 payments they analyzed were between one-hundredth of a cent and 10 cents — values too low for conventional credit-card rails.

Despite the growing counts, the total dollar value remains modest: in July, x402 processed about 75 million payments totaling roughly $24 million over 30 days. Transaction counts alone do not prove broad commercial demand and prior research has raised concerns about artificial or noncommercial activity on the network.

Trends in Wallet Behavior

Chainalysis also examined funding and usage patterns. Among wallets first funded in Q2 2026, 89.3% received additional funding within 30 days, versus 70.6% for wallets funded in Q3 2025 (the analysis excluded wallets that made payments on only one day). Another striking shift is timing: the average interval between funding and the first x402 payment dropped from 73 days in late 2025 to just 18 minutes in Q3 2026.

Regional Nuances and Service Types

Activity patterns differ by geography and service category. For example, Chinese-linked wallets represented roughly 90% of geographically identified payments for security and code-analysis services, much of it paid to AgentLisa, an AI-driven smart-contract auditing platform. Chainalysis warns against overinterpreting this as evidence that open-source models are eroding Asia’s on-chain inference market; multiple factors could explain the divergence.

Notably, x402 activity measures only identifiable transactions through that protocol, not the entire global market for AI-related payments.

Implications

As AI agents take on economic roles, each agent effectively becomes a new blockchain ‘user.’ While current transaction sizes are small, these behaviors hint at the emergence of autonomous spending patterns led by the United States in paid AI services. If trends continue, micropaids via stablecoins and protocols such as x402 may become an essential layer of agentic commerce.