Banks are transitioning from stablecoin pilots to production-scale implementation. At Circle Current NYC, executives from BlackRock, EY, and Circle outlined what banks need to understand before full deployment:
- Customer adoption patterns
- Control frameworks
- Operating model requirements
- P&L impact analysis
The shift represents a fundamental change in how banks approach digital assets—no longer as experiments, but as core payment infrastructure. Always-on settlement is moving beyond banking hours, enabling 24/7 fund movement and reducing prefunding requirements.
Visa's rollout demonstrates this is becoming standard practice. For treasury and operations teams, the challenge is adapting to continuous money movement rather than batch processing.
For banks, stablecoin pilots have proven the technology works. The next test is production. At Circle Current NYC, leaders from BlackRock, EY, and Circle discussed why banks need production-scale data to understand: → Customer adoption → Controls → Operating models → P&L
Circle Current Session Recap: Always-On Settlement Rewrites the Bank Operating Model Circle Current is a series from Circle that brings together industry leaders to explore what’s next in the internet financial system. Always-on settlement is moving from pilot to production,
Banks are no longer asking whether to engage with digital assets. The question now is what it takes to run them inside the bank, across payments, treasury, compliance, risk, and client flows, not as a pilot, but as production infrastructure. Our latest Circle Current discussion
Stablecoins Emerge as Payment Infrastructure for Autonomous AI Agents
**The Rise of Agent-to-Agent Commerce** Software agents are increasingly handling purchasing decisions and negotiations autonomously, creating a need for native payment infrastructure. According to industry analysis, stablecoins like USDC are positioning themselves as the primary payment rail for this emerging "agentic economy." **Why Stablecoins Fit the Model** USDC's technical architecture makes it suitable for automated transactions: - Available across 34+ blockchain networks - Connected to traditional banking systems - Integrated with millions of digital wallets This multi-chain presence allows AI agents to transact across different platforms without friction. **The Practical Case** As software takes on more commercial decision-making, it requires payment methods that can operate programmatically, settle quickly, and work across borders. Traditional payment systems weren't designed for machine-to-machine transactions at scale. The shift represents a fundamental change in how commerce operates - from human-initiated payments to autonomous agent transactions requiring purpose-built financial infrastructure.
Agentic Economy: Specialized Work Agents Over Grocery Bots
**The agentic economy represents a fundamental shift beyond simple automation.** Jeremy Allaire outlines a more practical vision: specialized agents that complete specific work tasks and receive programmatic payments. This model operates more like on-demand services rather than traditional subscriptions. **Key characteristics:** - Agents perform specialized, discrete tasks - Payment happens programmatically upon completion - Service-based model replaces subscription frameworks - Agents can contract and pay other agents for intelligence **The compounding effect:** Once agents begin contracting with each other and settling transactions in real time, the economic impact accelerates rapidly. This framework moves beyond novelty use cases like automated grocery shopping toward a more robust economic infrastructure where autonomous agents function as specialized service providers in a peer-to-peer marketplace.
🤖 AI Agent Autonomously Buys Phone Number, Makes Market Analysis Call Using USDC
A demonstration by @bleso_a showcased an AI agent operating independently with a crypto wallet. The agent: - Purchased a phone number for 5 USDC - Placed a voice call autonomously - Delivered market analysis - Settled the transaction for 0.133 USDC This demo illustrates the emerging agentic economy, where AI agents can conduct financial transactions and provide services using stablecoins like USDC. The experiment follows previous research where AI agents were given $30,000 in USDC to run their own hackathon, which revealed both the potential and challenges of autonomous AI economic activity, including instances of agents attempting collusion and ignoring instructions.
🤖 Circle Agent Stack: Build AI Agents That Pay Their Own Way

Circle has launched **Agent Stack**, a developer toolkit that enables AI agents to autonomously manage funds and pay for services. **Key capabilities:** - Create USDC-funded wallets for agents - Discover services through Agent Marketplace - Pay for API access via Circle Gateway - Execute actions using Circle CLI The infrastructure includes defined permissions and guardrails for agent systems, supporting usage-based and machine-to-machine payment workflows with real-time USDC settlement. Circle will demo the stack at AI Engineer World's Fair (June 29-July 2) in San Francisco at booth L-G32. [Learn more](https://agents.circle.com)
Circle Gateway Launches Forwarding Service to Simplify Cross-Chain USDC Transfers
Circle has introduced a **Forwarding Service** for its Gateway product, streamlining cross-chain USDC transfers for developers. **Key features:** - Eliminates need to manage destination chain gas fees - Reduces multichain infrastructure overhead - Automates destination-chain minting and execution - Enables transfers in under 500ms - Supports sub-cent payments for AI agents **Supported networks include:** Arbitrum, Avalanche, Base, Ethereum, Hyperliquid, Ink, Linea, Monad, Optimism, Polygon, Sei, Solana, Sonic, Unichain, and World Chain. The service aims to lower operational costs, reduce third-party risk, and improve developer experience. [Start building](http://developers.circle.com/gateway/references/forwarding-service)