When a user tells an AI assistant, "Book me a flight to Beijing for tomorrow," the AI automatically handles searching, price comparison, booking, and payment—something that, even in 2026, still feels like science fiction. Yet the real bottleneck isn’t logical reasoning, but the payment process: when an AI agent needs to pay $0.05 for an API call, traditional payment networks charge a $0.30 fee. From any perspective, this makes the transaction unviable, or even impossible to process.
At the World Economic Forum in Davos, Circle CEO Jeremy Allaire bluntly stated that traditional payment systems are "too slow, too expensive, and too cumbersome" for the emerging world of AI agents. When hundreds of millions of smart devices need to make automated payments, only blockchain networks based on stablecoins can support such massive transaction volumes—because they are the only financial infrastructure that delivers instant settlement, ultra-low costs, global accessibility, and price stability all at once.
This isn’t a distant future scenario. Between May 2025 and April 2026, AI agents completed approximately 176 million on-chain transactions across multiple blockchain networks, settling over $73 million in total. The average payment per transaction was just $0.31 to $0.48. The programmability, low-latency settlement, and global liquidity of crypto assets make on-chain infrastructure the natural choice for autonomous financial operations by AI agents.
AI Agent Payments Are Reaching a Structural Inflection Point
In May 2026, crypto market maker and investment firm Keyrock, together with Coinbase, Tempo, and Virtuals Protocol, released the "Who Pays the Agent?" report, the first systematic presentation of large-scale data on AI agent payments.
According to the report, about 76% of AI agent transactions fall below Visa’s fixed fee threshold of $0.30, with 98.6% of payments settled in USDC. As of Q1 2026, over 104,000 AI agents had completed registration. On the Base network, a single USDC transfer costs about $0.0001—just 0.03% of a $0.31 transaction. This isn’t a marginal cost improvement; it’s the fundamental reason for structural replacement.
On a broader scale, global stablecoin transaction volume reached $28 trillion in Q1 2026, with approximately 76% driven by automated systems and bots. At the same time, retail transfers dropped by 16%—the largest decrease on record. This means that machine-to-machine payments are no longer a fringe use case for blockchains; they are now a core force driving the transformation of the entire payment system.
Why Can’t Traditional Payment Systems Meet AI Agent Payment Needs?
An AI agent programmed to monitor on-chain arbitrage opportunities and execute trades cannot achieve full autonomy if it can’t pay transaction fees, access paid APIs for real-time data, or settle service fees with other agents.
Traditional payment systems were never designed for programmatic entities. Bank accounts require human identity verification, payment confirmations need SMS or biometric checks, and batch settlements face strict compliance reviews. When an AI agent needs to pay $0.05 for a single API call, traditional card networks can’t even process the request. These aren’t minor technical issues—they’re fundamental structural flaws of legacy payment systems.
In contrast, stablecoin-based on-chain payments offer an entirely different cost model. The x402 protocol, inspired by the HTTP status code "402 Payment Required," deeply integrates payment capabilities into web HTTP requests, enabling AI agents, APIs, and various applications to complete transactions simultaneously with service requests. In May 2026, the Linux Foundation officially established the x402 Foundation to promote this standard through open-source development, with members including Amazon, Google, Microsoft, Mastercard, Visa, Shopify, and other industry leaders. This marks the shift of machine autonomous payments from experimental projects to industry-standard consensus.
The Industry Infrastructure for AI Agent Payments Is Rapidly Taking Shape in 2026
Since 2026, tech giants and crypto companies have been launching infrastructure products specifically for AI agent payments.
In May 2026, AWS announced Amazon Bedrock AgentCore Payments, a system developed in partnership with Coinbase and Stripe. This platform enables AI agents to use stablecoins for automated payments to APIs, data sources, web content, and other online services during task execution. Developers can choose wallets from Coinbase or Stripe to fund agents with stablecoins, allowing them to make micropayments as low as $0.01, with real-time transaction support.
Around the same time, the Solana Foundation and Google Cloud jointly launched Pay.sh, allowing AI agents to use stablecoins on Solana for pay-per-use access to Google Cloud APIs and APIs from over 50 other providers. Pay.sh reduces the cost of a single API call to under one cent, enabling agents to autonomously discover APIs, check pricing, and complete payments directly. Other key players include Stripe and Tempo’s Machine Payments Protocol (MPP), Google’s AP2 delegated payment system, and Visa’s expanded tokenized payment credential service.
Two Distinct Technical Approaches
The AI agent payment landscape has now split into two clear paths. One camp relies on tokenized card credentials controlled by Visa and Mastercard to build payment rails. Mastercard launched its intelligent agent payment service in April 2025, partnering with Microsoft, IBM’s watsonx, and others. The other camp is led by the stablecoin ecosystem, using open protocols like x402 to settle directly with stablecoins over the internet.
Each approach has its strengths. Card-based channels excel in consumer retail scenarios, with robust mechanisms for chargebacks, fraud prevention, and dispute resolution. Stablecoin channels, however, offer structural cost advantages for high-frequency, low-value, cross-border machine transactions—scenarios where traditional card fees and settlement times simply don’t work. Going forward, key factors shaping the market will include not just fee differences, but also the development of agent identity verification and dispute resolution mechanisms.
AI Agent On-Chain Network Scale Is Expanding Rapidly
Real market data confirms this growth. The number of unique AI agent addresses with stablecoin activity on the Base chain has surpassed 12,500. In the past 30 days, agents initiated stablecoin transfers totaling $4.7 billion—about 320% higher than the quarterly average. The share of agent-to-agent direct transfers among all agent stablecoin transactions jumped from 15% at the end of 2025 to 38% in May 2026—evidence that network effects among agents are self-reinforcing.
On BNB Chain, the number of deployed AI agents soared from fewer than 400 in January 2026 to over 150,000 by April 2026—a staggering 43,750% increase. Assuming each agent continuously schedules tasks that require payments, BNB Chain alone could see between 470 million and 1.1 billion orders settled on-chain.
In Q1 2026, global crypto trading volume reached $20.57 trillion. Of this, AI agent-driven activity accounted for over 15% of decentralized exchange volume—a significant jump from 3% a year earlier. Autonomous or AI-driven transactions now make up 19% of all on-chain activity, and this share continues to rise rapidly.
Gate for AI Agent: Delivering a Complete Technology Stack for Machine Autonomous Payments
Against this industry backdrop, Gate has built a full-stack infrastructure platform for AI agents. Gate for AI Agent uses a four-layer architecture, abstracting from the infrastructure layer up to the application layer, enabling AI agents to naturally access comprehensive crypto capabilities. The infrastructure layer includes exchanges, DEXs, wallets, information services, on-chain data, and native payment gateways. The protocol layer offers MCP, CLI, x402 payment protocol, and A2A agent communication protocol. The capability layer packages over 40 prebuilt AI Skills, covering market research, trade execution, asset management, on-chain interactions, and information delivery.
On the payment side, Gate has launched the Gate Pay for AI module, which leverages the x402 protocol to provide payment and settlement capabilities to AI agents in a structured way. Users can interact through a natural language interface, using Gate Skills and CLI to trigger underlying token transactions, on-chain interactions, and payment functions.
For security, Gate employs a "permission isolation and safety guardrail" mechanism: public query operations require no authorization, while sensitive actions like fund transfers and trade orders mandate secondary confirmation. Users can also create dedicated subaccounts for AI agents, isolating operational risk within a controlled environment.
Conclusion
As AI agents gain the ability to autonomously seek services, negotiate prices, and complete payments and settlements, economic activity will evolve from human-driven to agent-driven. Chainalysis market data shows that by the end of 2026, AI agents could account for up to 30% of on-chain activity. On Layer 2 networks, about 40% of stablecoin transfers are already driven by automated systems. Circle predicts that in the next five years, billions of AI agents will use stablecoins for autonomous operations and settlements.
However, this evolutionary path still faces real-world challenges. Keyrock’s report notes that the current AI agent payment ecosystem is highly dependent on USDC, posing centralization risks. The entire emerging payment system relies heavily on the regulatory and infrastructural stability of a single stablecoin issuer. In addition, regulatory frameworks such as the European MiCA, the US GENIUS Act, and the EU AI Act still lack comprehensive standards for AI-driven financial transactions and machine-to-machine payments.
Another practical bottleneck is identity verification. When a software program initiates a payment, merchants need to confirm that the operator is a legitimate agent acting on behalf of a real user, not a malicious bot using stolen credentials. Visa reports that AI traffic to US retail websites has surged 47-fold, and has partnered with Cloudflare to launch a trustworthy agent protocol to distinguish legitimate AI programs from malicious crawlers.
As technical standards, regulatory rules, and security verification mechanisms continue to mature, stablecoins and on-chain payments are becoming the foundational settlement layer for the AI agent economy. The infrastructure provided by Gate for AI Agent will further lower the barriers for AI agents to participate in the crypto economy, enabling seamless machine-to-machine economic activity. As more agents join this network, the growth curve of the on-chain economy will shift from linear accumulation to network-effect-driven acceleration.
