Autonomous AI agents are spending money. A lot of it. On Base, Coinbase’s Layer 2 network, agentic payments have crossed 20 million transfers in a single 90-day window, and the cumulative picture is even more striking: roughly 169 million total agentic transactions processed as of July 2026.
What agentic payments actually are
Agentic payments are transactions initiated by AI agents, not people. Think of software that autonomously pays for API access, data feeds, or compute resources, settling in stablecoins in seconds via a single HTTP request. No invoice, no accounts payable department, no waiting.
The infrastructure making this possible on Base is largely the x402 protocol. It’s named after HTTP status code 402, which was originally reserved for future payment functionality back in 1998 and never actually used. Someone finally used it.
In the 30 days leading up to May 29, 2026, x402 logged 3.1 million transactions with a combined transfer value of $1.2 million. Those are small-dollar, high-frequency payments, exactly the kind of micro-settlement pattern that’s impossible to run efficiently through traditional financial rails.
By June 3, 2026, cumulative agentic transactions on Base via x402 had crossed 100 million, according to a Chainalysis report. That number has since climbed to approximately 169 million as of July 2026, which means the back half of that growth happened fast.
The quality shift that matters more than the volume
In early 2025, transactions of $1 or more represented just 49% of Base’s total agentic transfer volume. As of the most recent data, that figure has risen to 95%.
The early growth on Base was partly driven by meme coin activity and low-value automated farming. The trajectory since then has moved toward utility.
Base currently holds over $4 billion in Total Value Locked.
Why Stripe and Google showing up matters
Stripe integrated in February 2026 and Google followed, with both companies involved in standardizing agentic commerce infrastructure. The x402 protocol’s design, settling stablecoin payments for digital resources like APIs and data at low cost in seconds, maps directly onto what AI agent workflows actually need.
What investors should be watching
Regulatory attention is the variable most likely to complicate this picture. Autonomous machine-to-machine payments represent a genuinely new category for regulators: who is responsible when an AI agent initiates a transaction? The current frameworks weren’t written with that question in mind.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.

1 hour ago
3














English (US) ·