Artificial intelligence (AI) agents spending money online autonomously is still a tiny market, but some of the world’s biggest tech, payments and crypto companies are already racing to build the necessary infrastructure, Keyrock said in a new report.
The crypto trading and investment firm estimated that AI agents settled more than $73 million on approximately 176 million transactions on the blockchain between May 2025 and April 2026.
Volumes remain negligible compared to traditional finance (TradFi). Visa, for example, alone processes $14.5 trillion per year. But the importance lies less in the overall value of the U.S. dollar than in how quickly the infrastructure stack forms, the report asserts. Global companies such as Coinbase (COIN), Stripe, Google (GOOG), and Visa (V) have all deployed competing systems for machine-to-machine payments.
The broader idea behind agent payments is that software increasingly consumes digital services autonomously rather than through subscriptions and accounts managed by humans. An AI sales agent, for example, could continuously purchase market data, cloud computing, or AI-generated analytics in small increments throughout the day without a human manually authorizing each payment.
This potential gives rise to ambitious forecasts as to the scale that the agent payments sector could take. Gartner predicts that AI agents could make $15 trillion in purchases by 2028, while McKinsey estimates that retail agent commerce could reach $3 trillion to $5 trillion by 2030, according to the Keyrock report.
These projections imply even faster growth rates than stablecoins during their breakout years, the report says, but indicates that the pace of infrastructure deployment already indicates that the market is moving beyond its experimentation phase.
Coinbase’s x402 protocol has become one of the leading native cryptosystems. The protocol allows AI agents to pay directly with USDC for services such as blockchain analysis or cloud infrastructure without creating an account or subscription.
Stripe, with its Tempo blockchain, launched a competing framework called Machine Payments Protocol (MPP), while Google introduced AP2, a system focused on delegated spending authorization for AI agents. Visa has expanded its card network with tokenized credentials designed for AI-driven commerce.
Crypto rails and stablecoins are emerging as the preferred settlement layer, and economics helps explain why.
According to the report, about 76% of agent transactions fall below the 30-cent flat fee floor common in card payments. Most payments ranged from 1 to 10 cents, making traditional rails impractical for automated software agents purchasing data, AI inference, or API access. Meanwhile, settling stablecoins on some blockchains like Base and Tempo costs fractions of a cent.
Currently, 98.6% of automatic payments are settled in USDC, the stablecoin issued by Circle (CRCL). This strengthens Circle’s position in crypto payments, but also introduces concentration risk, creating dependence on a single issuer.
Regulation could be a source of constraints on growth. Europe’s MiCA, the U.S. GENIUS Act, and the EU AI Act are all expected to take effect around mid-2026, but none of them directly address autonomous machine-to-machine transactions or questions about agent liability and identity, the report notes.
