Coinbase Launches Batch Settlement for AI Agent Payments

The Coinbase-backed x402 protocol launched batch settlement to support high-frequency, low-cost payments between autonomous AI agents.

By Daniel Mercer Published:

The Coinbase-backed x402 protocol has launched batch settlement functionality designed to support high-frequency payments between autonomous AI agents, marking a significant step toward scalable machine-to-machine commerce.

The new system allows agents to authorize multiple off-chain micro-payments and settle them later in bulk on-chain, dramatically reducing transaction overhead. According to Base creator Jesse Pollak, the feature enables payments as small as $0.0001, a threshold considered critical for the emerging agentic economy.

Under the system, buyers deposit ERC-20 funds into escrow accounts while agents sign cumulative off-chain payment vouchers for individual requests. Sellers can instantly verify these vouchers and deliver services without waiting for blockchain confirmation, while settlements are later grouped into a single on-chain transaction.

The protocol is aimed at supporting AI-driven demand for compute power, model inference, APIs, and other digital resources that require rapid, low-cost transactions. Coinbase Developer Platform product lead Joshua Nickerson said the structure creates a near-gasless experience for AI agents by shifting transaction costs to facilitators.

The launch follows recent integration of x402 infrastructure by Amazon Web Services, enabling AI agents to make USDC payments on Base and Solana without direct private key management.

The development comes amid intensifying competition to build financial infrastructure for autonomous AI systems. Companies including Circle and organizations such as Aptos Foundation are also investing heavily in AI-focused payment rails and programmable financial systems.

The rollout highlights a growing convergence between artificial intelligence and blockchain infrastructure, where scalable micro-transactions are increasingly viewed as essential for enabling autonomous digital economies and machine-native commerce.

AI & Machine Learning, News