Topics Pioneers Insight Method Author
Back

Is Circle More Than Just USDC? Arc, AI Agents, and What’s Next

·The Edge Podcast

Rendering the long image…

Topics Synthetic Dollars and Crypto Neobanks Crypto Payments for AI Agents
Pioneers Gagan Mac

Summary

  • Circle’s broader bet is that programmable dollars can support an entire financial infrastructure stack. Gagan describes a “neutral market infrastructure company” serving enterprises, developers, and agents through USDC, payments infrastructure, interoperability, and Arc. He says USDC crossed $100 trillion in transaction volume “last week” and now operates on 30-plus networks.
  • USDC’s transfer volume reflects software-driven reuse of money as well as payment adoption. Nomadic cites potentially inaccurate, LLM-sourced 2025 figures of $18.3 trillion for USDC transfers versus Visa’s $14.2 trillion and Mastercard’s $10.6 trillion. Gagan’s explanation is that “each dollar that is on the internet makes multiple turns”: settlement can immediately trigger lending, collateral movements, or another trade.
  • Gagan’s defense against competing stablecoins rests on distribution and utility that issuance alone cannot reproduce. He says USDC has one-third of Tether’s circulation but transacts more, supported by trusted reserves, cross-chain connectivity, and developer adoption. Neobanks can obtain branding and reserve economics through Circle partnerships; the harder question is, “How will you get the liquidity for it, the distribution for it?”
  • Arc targets institutional settlement while Circle continues to support multiple blockchains. Its proposed advantages are institutional validators, USDC-denominated gas, and “sub-second finality,” addressing enterprise compliance and accounting requirements. Gagan explicitly rejects the expectation that all Ethereum liquidity migrates to Arc, describing “coexisting and interconnected networks.”
  • Agent payments are a major growth bet with substantial engineering still required. Gagan says 99.6% of agentic transactions use USDC and expects transaction volume to grow 100× from here. Circle’s x402 facilitator and nanopayments address machine purchases, but micro-cent transactions cannot work with cent-sized fees: “there are still a lot of problems to solve.”
  • Circle’s next priorities extend from confidential transactions to agent trust and broader validator participation. Gagan highlights privacy compatible with existing smart-contract structures, a “chain of trust” built around auditable agent work, and a full proof-of-stake launch “with our token.” He wants additional validators and greater network resilience over the next year; the agent trust architecture remains “a very interesting bet.”

Deep dive

1. Circle wants to make finance programmable

  • Gagan traces his conviction to Twilio: a few lines of code let him broadcast messages for his sister’s 2,000-person wedding in India. Programmable communications enabled authentication, receipts, and notifications. He wants the same transformation for money: use the existing regulated dollar system as a backbone and “make the dollar digital-native,” so businesses can build around it.

  • Circle’s mission is “to enable frictionless exchange of value globally.” Gagan says USDC crossed $100 trillion in transaction volume “last week” and expanded to 30-plus networks over five years. Circle Payments Network supplies infrastructure for compliant cross-border payments: fiat on-ramps, stablecoin settlement, RFQ/FX systems, and the receiving fiat leg. His description of Circle is “a neutral market infrastructure company.”

  • The North Star is equal access to financial tools, analogous to email or WhatsApp working for anyone online. Local builders should give users access to global assets, including US equities, Treasuries, or blockchain representations. Arc is the proposed “economic OS for the internet”; digital assets sit above it, with payments as one application serving enterprises, developers, and agents.

2. Transaction volume measures more than consumer payments

  • Nomadic’s scale comparison comes with an explicit caveat: “This might not be exactly right,” because his figures came from an LLM. For 2025, he cites $18.3 trillion of USDC on-chain transfers against Visa’s $14.2 trillion and Mastercard’s $10.6 trillion, or $24.8 trillion combined. The comparison makes outsiders say “No way”: stablecoins have reached a scale many still do not recognize.

  • Gagan says stablecoins overall have already leapfrogged card-network transaction volumes, while acknowledging those networks’ greater physical retail reach. The mechanism is broader utility: payments, collateral, and trades triggered by incoming funds. “Each dollar that is on the internet makes multiple turns.” Card payments can remain pending for 48–72 hours; stablecoin settlement can immediately trigger another transaction, letting software and AI bots reuse money faster.

  • For competitive positioning, Gagan says USDC has one-third of Tether’s circulation yet transacts more. He attributes that to trusted, audited reserves and programmability, reinforced by applications such as Hyperliquid and Polymarket on different networks. CCTP, his “TCP/IP for moving money,” moves USDC across 30 networks so developers can build wherever they choose while users retain a familiar currency.

3. Liquidity is the barrier new stablecoins must clear

  • Nomadic asks whether Circle must pick winners, pointing to roughly 63% of USDC on Ethereum and around 9% on Hyperliquid, plus the “aligned quote asset V2” arrangement. Gagan rejects that premise: “We didn’t want to be in a position where we were trying to pick winners.” Circle worked with Hyperliquid for almost a year on its migration from Arbitrum to HyperEVM, targeting a better user experience, more security, and new experiences on its network.

  • Demand grew from four or five supported networks roughly two and a half years ago to 40–50 requesting stablecoins. The initial “Bridge to Native” standard let networks wrap USDC, but Gagan says the user experience was poor. Circle then built a scalable deployment process: meet compliance, user-safety, and security criteria, align partnership incentives, and deploy USDC on the network. Its stated objective is to support more builders without selecting ecosystem winners.

  • When neobanks request their own stablecoin, Gagan first asks what they want. Earning the Treasury rate on reserve assets does not require issuance: “You don’t need to build your own stablecoin for that.” Circle offers distribution partnerships and an existing program for branded stablecoins on top of USDC, while retaining the economic benefits.

  • The harder requirement is giving customers something useful to do with the money. Gagan’s example is a Morpho vault where users can earn 6% or 7% by lending. That requires developers to borrow, lend, and accept the stablecoin as collateral, plus sufficient liquidity. “Creating the stablecoin is only the starting point.” He sees value in better earning opportunities and faster expansion into new regions.

4. Agents need payments small enough for individual tasks

  • For agents, the immediate bottleneck is access to paid tools: subscriptions usually require a card, bank rails, and onboarding. Gagan argues x402 lets one machine pay another directly with instantly settling stablecoins. Circle provides an x402 facilitator on Arc; he also notes facilitators on Solana and other networks. Agents can then purchase tools needed to complete their assigned goals.

  • Nanopayments targets micro-cent purchases by batching very small payments while keeping costs minimal. Gagan’s constraint is straightforward: “If you’re making a payment of a microcent and paying a cent in network fees, that doesn’t work.” Circle is investing in faster payments and higher throughput alongside lower costs; “there are still a lot of problems to solve.”

  • The scale thesis is explicit: Gagan says 99.6% of agentic transactions use USDC, attributing that to liquidity, programmability, and network effects. At scale, he envisages machines making 100,000 times as many payments as humans and expects transaction volume to grow 100× from here. Those expectations drive investment in the full agent stack.

5. Arc’s architecture starts with enterprise settlement requirements

  • DeFi Dad challenges the architecture: why build an L1 rather than inherit Ethereum’s security through an L2? Gagan starts with enterprise risk teams needing identifiable trust assumptions. His illustrative validator pitch names BlackRock, SBI, DTCC, Mastercard, and Visa—institutions a compliance team recognizes. The desired design is distributed control by global institutions; he says no other network was able to offer that proposition.

  • USDC as gas addresses another enterprise problem: volatile tokens create budgeting and accounting complexity. Arc also targets “subsecond finality” using BFT consensus. Gagan contrasts this with L2s periodically settling batches and the possibility of reorganizations; his claimed benefit is knowing payment is settled within a second. Circle contemplated an L2, then moved toward an L1 as these requirements accumulated.

  • Nomadic presses whether Arc should capture all USDC volume and liquidity. Gagan rejects that outcome: “We absolutely don’t” envision all Ethereum liquidity migrating. Ethereum’s battle-tested distributed architecture remains important, and Circle will keep building protocols there alongside Arc. His intended relationship is “coexisting and interconnected networks,” with purpose-built chains serving different financial use cases.

  • That position has a concrete interoperability component: Gagan says Circle opened its interop stack “yesterday” to Euro Coin and third-party asset issuers, extending connectivity already used by USDC across Ethereum, Solana, Base, Arc, and other networks. Arc also has sectors for payments and agents with differing requirements. Networks, in his framing, serve the utility delivered to builders and users.

6. Privacy and agent trust lead the next product agenda

  • The first priority is Arc privacy, previewed “yesterday.” Gagan wants confidentiality for asset issuers and customers without breaking smart-contract structures or EVM compatibility. Once an address is identified, analysts can track its holdings. He thinks privacy unlocks payroll and novel trading use cases that don’t happen in public.

  • The second is an agent trust stack. Treating agents as workers raises the question of whether personal or banking data will leak. Circle is investing in secure processing guarantees and hashed on-chain audit logs so an agent’s work can be audited. Gagan calls the broader idea a “chain of trust,” where tamper-proof guarantees on the network could support trust in agents.

  • His analogy is an Airbnb host with 100,000 reviews and five stars: accumulated history helps a stranger decide whether to trust the service. For agents, a longer auditable work history could provide a similar signal. Gagan calls this “an interesting bet and idea” and believes Circle can extend trust established through open dollar software to AI on the open internet.

  • The third priority is progressing from foundational institutional validators toward “a full proof-of-stake launch” with “our token.” Gagan wants to onboard more validators and make Arc more resilient over the next year.