NEAR Protocol has launched private payments for agentic commerce, adding confidential USDC transactions to its AI-focused infrastructure. The feature uses Confidential Intents to let AI agents transact without exposing sensitive financial details on public blockchain rails.
The launch connects Circle’s USDC with NEAR’s agent infrastructure, giving autonomous software a stable payment asset for on-chain commerce. The goal is to move AI agents beyond recommendations and into controlled financial execution.
Confidential Intents Target Agent Payment Privacy
The system addresses a core limitation of public blockchain payments. While open ledgers provide transparency, that same visibility can expose transaction behavior, payment amounts and commercial relationships that businesses or individuals may need to keep private.
Confidential Intents are designed to hide sensitive payment data while preserving settlement functionality. That structure makes the payment layer more suitable for agentic commerce, where software agents may need to purchase services, access tools or execute transactions on behalf of users.
NEAR’s AI stack brings together private inference, agent identity and cross-chain settlement infrastructure. NEAR AI Cloud supports private and verifiable inference, Ironclaw manages agent deployment and policy enforcement, and NEAR Intents handles discovery, negotiation and execution.
Agentic Commerce Needs Governance and Control
The launch reflects a broader shift toward programmatic payments between software systems. AI agents will need ways to pay for compute, data, APIs, subscriptions and digital services without relying on manual user approval for every small transaction.
That creates new requirements around permissions, spending limits and financial privacy. If users delegate payment authority to agents, the infrastructure must define what agents can spend, where they can transact and how activity can be reviewed or restricted.
NEAR’s recent involvement with the x402 Foundation also fits this direction. The foundation is working on internet-native payment standards for AI agents, while NEAR is positioning Confidential Intents as part of the privacy layer needed for that economy.
The key test is whether private agent payments can remain usable, auditable and secure at scale. Enterprise adoption will depend on reliability, policy controls, stablecoin liquidity and confidence that confidential settlement does not weaken user protection.
NEAR’s private USDC payments give the agentic commerce stack a live privacy-focused settlement feature. The next useful indicators will be agent transaction volume, developer adoption, USDC payment flows, Ironclaw policy usage and whether real applications begin using Confidential Intents for recurring automated commerce.