DEV Community

Cover image for AI Agents Settle $280M in Micro-Payments: Solana Captures 74% as Base Fees Spike
Fin Pulse
Fin Pulse

Posted on Originally published at finpulse360.org

AI Agents Settle $280M in Micro-Payments: Solana Captures 74% as Base Fees Spike

AI Agents Settle $280M in Micro-Payments: Solana Captures 74% as Base Fees Spike

2026-11-01, FP360 — Autonomous AI agents processed $280M in micro-transactions over the past 30 days, with Solana capturing 74% of volume as Base network fees spiked to $0.18 per tx, driving agent activity to Solana's $0.0004 average. The surge, up 40% month-over-month, is fueled by agent-to-agent API payments and DePIN compute billing, pressuring Ethereum L2s to cut costs or lose the next wave of machine-to-machine commerce.

The Signal: Solana's fee advantage and high throughput are pulling AI agent micro-transactions away from Base, a trend likely to accelerate until Base deploys its next fee upgrade.


Originally published on FINPULSE 360

Top comments (1)

Collapse
 
tercelyi profile image
tercel •

The 74% number basically screams “fee-sensitive, high-frequency flow will go wherever it’s cheapest and ‘good enough’.” At $0.18 on Base vs ~$0.0004 on Solana, AI agents aren’t making a philosophical choice, they’re doing strict unit economics.

At $280M over 30 days and mostly micro-payments, a few questions pop up:

  • What’s the median tx size and count per agent? If agents are firing thousands of calls per day, even a $0.01 fee would be huge friction.
  • How much of that 74% is sticky vs purely opportunistic routing? Are devs integrating Solana-first, or is it just the current cheapest RPC endpoint in a multi-chain setup?
  • For DePIN compute, is latency or finality starting to matter as much as fee, or is cost still the overwhelming driver?

The pressure on L2s feels under-acknowledged. An “AI agent M2M lane” might be a separate design space from human-facing DeFi / NFT flows. For agents, UX = predictable, near-zero-cost, composable APIs, not wallets and frontends.

Curious what you’re seeing in the data: do failed / dropped txs on higher-fee chains spike right before flow migrates, or is the switch more gradual as devs update routing logic?