138 million AI agent payments on the Base chain highlight micro-payment growth

Written by Xiao Bing

A status code forgotten by the HTTP protocol for thirty years is becoming the cash register of the machine economy.

Chainalysis's North American Crypto Market Report, released on October 14, revealed a set of data: as of September 2026, over 138 million AI Agent payment transactions have been completed on Coinbase’s Base chain via the x402 protocol. Including other chains such as Solana, the total transaction volume in the x402 ecosystem has surpassed 165 million, with payment amounts exceeding $50 million.

In these transactions, 87% of individual amounts fall between 0.01 cent and 10 cents. The most common payment range is 1 to 10 cents.

Think of it this way: about five on-chain payments from AI agents are happening every second, each costing less than the price of a lollipop.

The story begins with HTTP 402.

In 1997, the HTTP/1.1 standard defined a status code "402 Payment Required," envisioning that future web services could use it to require users to pay before accessing content. However, at the time, no payment infrastructure existed to support microtransactions at the web page level, so this status code was marked as "reserved for future use" and remained unused for over two decades.

In May 2025, Coinbase picked it up. The x402 protocol is designed very simply: when an AI Agent sends a request to an API, if the server requires payment, it returns a 402 status code along with the price, payment address, and supported tokens. The Agent’s wallet automatically signs and settles the transaction on-chain to gain access. The entire process requires no account registration, no API key, and no human approval.

An HTTP request, a USDC transfer, a data exchange. Three steps to complete a machine-to-machine business transaction.

This mechanism works because stablecoins have driven settlement costs low enough. With Visa, processing a $0.03 transaction might cost more in fees than the transaction itself. But settling with USDC on Base costs less than one cent. When individual transactions can be as small as one-hundredth of a cent, the entire "pay-per-use" business model becomes viable.

Chainalysis's report breaks down the consumption scenarios behind these 138 million transactions, revealing several categories:

Model inference is the largest single category. AI agents call large language model APIs and pay per token. A single Q&A may cost $0.02, while code generation may cost $0.08. U.S. wallets account for 88.1% of model inference payments, consistent with the market position of leading U.S. model providers (OpenAI, Anthropic, Google).

Data and search APIs will follow soon. Agents pay for real-time data: price feeds, on-chain analytics, news summaries, and academic paper searches. The geographic distribution of these payments is more balanced, with U.S. wallets accounting for 23.2%.

Code auditing and security analysis was an unexpected highlight. Wallets from China accounted for approximately 90% of the transaction volume in this category, led by AgentLISA, an automated smart contract auditing service. Users submit contract code and pay per audit in USDC to receive reports.

Financial intelligence and market data are also growing rapidly. Chinese wallets account for 46.7% in this category, while U.S. wallets make up a relatively small share. This may reflect a reality: Chinese AI application developers are using overseas data services to bridge information gaps.

An implicit geographic divide has emerged: U.S. Agent usage favors the "production side" (reasoning, generation), while China's Agent usage leans toward the "analysis side" (auditing, intelligence). Chainalysis attributes China's extremely low share in the reasoning category (0.1%) to "AI nationalism": domestic open-source models such as DeepSeek and Qwen are freely available, eliminating the need to pay on-chain fees to access overseas commercial APIs.

More noteworthy than trading volume are changes in user behavior.

A significant portion of x402's early growth came from PING, a meme coin that required payment in x402 to mint. This wave of speculation brought a surge of wallets, but also led to a subsequent decline in traffic. However, after the tide receded, the remaining users exhibited a completely different behavior pattern.

Chainalysis tracked the wallet's "stickiness metric":

The percentage of wallets that recharged within 30 days of their first deposit increased from 70.6% in Q3 2025 to 89.3% in Q2 2026. Nearly nine in ten wallets return to use the platform within a month of their initial deposit.

Even more striking is the change in "time to first spend." By the end of 2025, the median time from wallet creation to the first x402 payment was 73 days. By Q3 2026, this number had dropped to 18 minutes.

The percentage of wallets that began spending within one hour after depositing surged from 4% to 67%.

These three sets of data together tell the same story: x402 is transitioning from a "new toy for developers to experiment with" to the "default payment layer in Agent workflows." Wallets are no longer exploration tools driven by human curiosity, but rather programmatic infrastructure components. Funds are deposited and spent automatically, operating like utility bills that are deducted seamlessly.

The structure of payment amounts is also shifting. At the beginning of 2025, transactions over $1 accounted for only 49% of total payment volume; by the beginning of 2026, this share had risen to 95%. Transactions between $0.10 and $1 dropped sharply from 46% to 4%. Agents are no longer just making "trial小额 calls"—they are paying for heavier, more expensive services: renting GPUs by the hour, bulk-purchasing datasets, and subscribing to continuously updated financial data streams.

x402 is worth taking seriously not just because on-chain transaction volume is rising, but also because of the list of backers betting on it.

In May 2026, AWS released Bedrock AgentCore Payments, enabling AI agents running on Amazon Web Services to natively support x402 payments. Developers can choose a wallet from Coinbase or Stripe, fund it with USDC, and the agent will autonomously pay for API calls, MCP server access, and inter-agent collaboration.

Coinbase and AWS jointly launched the x402 Foundation, with Cloudflare as a founding member and native x402 support built into Workers. Google has integrated x402 into its Agent Payments Protocol (AP2). Visa has announced support for x402 through the Trusted Agent Protocol (TAP).

OpenAI and AWS have jointly released a technical guide for the x402 payment stream. The Solana Foundation has built a similar agent payment solution for Google Cloud services.

This list reads less like a roster of ecosystem partners for crypto projects and more like a consensus on infrastructure for the tech industry. When AWS, Google, Visa, and Stripe all appear together on a protocol’s list of supporters, the conversation is no longer about “whether cryptocurrency can go mainstream,” but rather “what are the payment standards for the machine economy?”

Circle’s statement in the report is intriguing: stablecoins are laying the rails for "machine-to-machine settlements," handling transactions so small that credit card networks have no interest in processing them. Will Su of BlackRock noted that AI agents could create an entirely new type of payment demand: continuous, round-the-clock, high-frequency, low-amount payment streams that barely exist in human economic activity.

What needs to be kept in perspective is that the total payment amount of $50 million, spread across 165 million transactions, averages just $0.30 per transaction. In the context of cryptocurrency markets, where daily trading volumes often reach tens of billions of dollars, this is a figure at the level of statistical noise.

Approximately 69,000 AI agents are active on Agentic.Market. Most of these agents are still at the stage of "being functional" but are not yet ready for large-scale commercial deployment.

But there’s a notable parallel here: x402’s growth trajectory closely resembles that of early DeFi. In 2019, DeFi’s TVL was under $1 billion, and few took it seriously. Three years later, locked value peaked above $180 billion. What matters isn’t the current absolute number, but the shape of the growth curve and the underlying structural drivers.

The number of AI agents is growing exponentially, and each agent needs to autonomously acquire and pay for external services. Traditional payment systems (credit cards, bank transfers, PayPal) were not designed for "machine-autonomous payments"—they require human authentication, accounts, and minimum transaction amounts. Stablecoins combined with x402 fill a gap entirely unaddressed by existing financial infrastructure.

By the time you read this article, thousands of AI agents may have already silently completed another round of microtransactions on Base. They don’t sleep, don’t eat, and don’t take weekends off. They simply follow their code’s instructions, purchasing data and compute power on-chain, one cent at a time.

This could be the first large-scale use case for cryptocurrency that does not rely on speculation.