Circle’s USDC Push: Arc, CPN and the Bet on AI Micropayments

Circle is trying to become more than the company behind USDC — it wants to own the rails, the blockchain, and maybe even the payments layer for AI agents

Circle is pushing USDC from a simple stablecoin business into a broader payments and infrastructure play, with its own network, its own chain, and a long-shot bet on machine-to-machine money.

• Revenue is up, but the old model still dominates: Q1 2026 revenue reached $694 million, yet about 94% still came from reserve interest income.
• More USDC is staying in Circle’s ecosystem: activity routed through Circle-controlled platforms jumped from 6% to 17.2% in one quarter.
• CPN and Arc are the big strategic bets: Circle is building a payments network and a Layer 1 blockchain to capture more fee-based revenue.
• AI micropayments are the far horizon: Circle’s Agent Stack is aimed at autonomous agents and nano-payments, but it’s still mostly an optionality play.

Circle’s latest numbers show a company in transition, but not a company that has fully escaped the business model that built it. The firm posted Q1 2026 revenue of $694 million, up 20% year over year, while adjusted EBITDA rose 24% to $151 million and the adjusted EBITDA margin hit 53%. Solid results, sure. But the catch is hard to miss: around 94% of revenue still came from reserve interest income.

For anyone new to the space, USDC is Circle’s dollar-backed stablecoin, designed to track the U.S. dollar at roughly 1:1. Circle earns most of its money by holding the reserves that back those tokens and collecting interest on cash and Treasuries. That works nicely when rates are high. It gets a lot less sexy when rates fall and the yield faucet starts dripping instead of gushing.

That is the core tension here. Circle is still a stablecoin issuer living mostly off interest income, but it clearly wants to become something bigger: a digital-asset infrastructure company with more durable, fee-based revenue. In other words, Circle wants to stop being just the fuel and start owning the engine, the road, and the toll booth too.

Circle’s numbers look better — but the model is still vulnerable

Operationally, the company is moving in the right direction. Realized revenue margin hit a record 41.4%, and Circle says more of its USDC activity is now happening on its own systems rather than leaking out through outside platforms. The report says activity routed through Circle-controlled platforms rose from 6% to 17.2% in a single quarter, while the share on external platforms fell to about 55%.

That matters because Circle’s strategy is really a form of vertical integration, which simply means owning more of the payment flow end to end instead of letting third parties capture the economics. Circle wants more activity to move through Circle Mint and the Circle Payments Network (CPN), where it can keep more of the revenue and control the customer relationship.

Tiger Research framed that shift as an improvement in Circle’s “profit quality” rather than just headline growth. That’s a fair read. Not all revenue is equal, and in crypto there’s a big difference between money earned from actual usage and money earned because macro conditions happen to be friendly. One is durable. The other is basically a gift from the bond market.

Circle still has some work to do before anyone should start calling this a finished transition. Net income fell about 15% year over year to roughly $55 million, pressured by higher stock-based compensation and more spending on infrastructure and research and development. So yes, the company is growing. No, it is not magically printing bottom-line profit just because the ecosystem looks busy on a dashboard.

Why reserve income is both powerful and fragile

Reserve interest income is the cleanest explanation for Circle’s current business strength and its biggest weakness. When the U.S. Federal Reserve keeps rates elevated, Circle can earn a handsome return on the assets backing USDC. When rates come down, the revenue stream shrinks. Simple as that.

That dependence makes Circle unusually exposed to macro conditions it cannot control. A stablecoin issuer tied to yield is not the same thing as a software company with recurring subscription revenue. It’s closer to a financial plumbing business with a very specific sensitivity to monetary policy. That’s fine when policy is helping. It’s a problem when it isn’t.

This is why Circle’s push into payments infrastructure matters so much. The goal is to create more fee-based revenue from transfers, settlement, foreign exchange, and network usage. Fee revenue is slower to build and harder to fake, but it’s also a lot more durable. No one gets rich building a serious financial network by clinging to one macro variable and hoping for the best.

CPN is Circle’s attempt to own more of the settlement stack

The Circle Payments Network (CPN) sits at the center of this strategy. Circle says participation in CPN has grown to 136 financial institutions, with annualized throughput reaching roughly $8.3 billion. The network is Circle’s attempt to move from “we issue the asset” to “we also provide the rails that move it.”

That’s the real prize. If Circle can get banks, fintechs, exchanges, and payment firms to use its infrastructure directly, it can capture more economics from every transaction. It also reduces dependence on third-party venues that may support USDC but don’t necessarily funnel much value back to Circle.

The company’s broader thesis is straightforward: if USDC becomes the default dollar token in crypto and beyond, Circle should not be content playing the role of passive issuer. It wants to be the network that settles the transfer, clears the payment, and eventually handles conversion too. Smart move? Yes. Easy? Not remotely.

Hyperliquid is a useful case study for USDC distribution

One of the more interesting signs of USDC’s reach is its adoption by Hyperliquid, the decentralized trading venue that made USDC an official base trading pair. Hyperliquid matters because it is not just another random logo on a partner slide. It is a genuine usage node for stablecoins, especially in trading and collateral activity.

According to the data cited, Hyperliquid’s total value locked (TVL) reportedly grew from $2 billion to $4 billion year over year, and at one point reached $6 billion. Tiger Research suggested that if that growth continues, Hyperliquid could eventually account for more than 10% of USDC circulation.

That’s a big deal because Circle says USDC circulation is around $77 billion. A large circulating supply only matters if the token is actively used. Distribution venues like Hyperliquid help turn USDC from a parked asset into a working asset. That’s valuable for Circle, even if the economics are not always as juicy as the company would like.

Still, some skepticism is healthy here. Adoption metrics can be misleading if they are not tied to actual fee-bearing activity. A venue can become a major USDC sink without necessarily becoming a major profit center. In crypto, “used a lot” and “makes money” are very different sentences.

Arc is Circle’s boldest bet — and the riskiest one

Then there’s Arc, Circle’s planned Layer 1 blockchain. Arc is being positioned as a settlement and execution layer for cross-border payments and on-chain FX — which means currency conversion done directly on blockchain rails. Circle is also building a product called StableFX for exchange and conversion use cases.

Why build a chain at all? Because Circle wants to control more of the stack, reduce dependence on other networks, and capture more of the economics around settlement and foreign exchange. That’s classic vertical integration. It also makes sense if Circle believes stablecoin payments are going to become core financial plumbing rather than just another crypto side quest.

The pitch for Arc leans heavily on the inefficiency of legacy payments. The report cites World Bank data showing average global remittance costs of 6.36% and bank transfer costs of up to 14.99%. Those are ugly numbers. They are also exactly the kind of numbers that make a programmable settlement layer look compelling.

Arc’s testnet reportedly saw 430 million cumulative transactions and 3.26 million transactions in 24 hours, with more than 100 institutions said to be participating, including BlackRock, HSBC, Visa, and Amazon Web Services. That sounds impressive, and in fairness it is the sort of institutional interest that can help a new chain gain credibility.

But let’s not get carried away by testnet fireworks. Transaction counts are not the same thing as revenue, and “participating” is not the same thing as “deploying production capital at scale.” Crypto has seen more than enough polished demos that turned out to be little more than expensive PowerPoint with a blockchain mascot attached. Arc may be real, useful infrastructure. Or it may be another corporate chain that looked amazing until the market asked, “Okay, but who’s actually paying?”

AI micropayments are the long game

Circle’s most forward-looking bet is its Agent Stack, designed for AI-agent and machine-to-machine micropayments using USDC. The idea is that autonomous software agents could eventually pay for data, compute, access, and services without needing slow, expensive, or heavily intermediated payment rails.

That is not a crazy thesis. If machine-to-machine commerce really expands, stablecoins could become a natural settlement layer. Software doesn’t care about banking hours, card fees, or the legacy system’s weird little rituals. It just wants to pay and get paid. USDC fits that use case better than most existing payment systems.

Still, this is a future-facing bet, not a current earnings engine. The report frames 2026 as infrastructure buildout, 2027 as regulatory normalization, and 2028 as commercialization for the AI stack. That’s a long runway. The company appears to understand that, too, noting that the AI layer may contribute less to near-term earnings than to a higher “platform premium” — a fancy way of saying investors may value the vision before the cash flow shows up.

That can work in public markets if the story is strong enough. It can also get people in trouble if the story outruns the business. As always, the spreadsheet gets the last word.

Regulation could help, but it won’t do the work for Circle

The regulatory backdrop may give Circle a tailwind. The GENIUS Act is mentioned as a possible catalyst for stablecoin adoption, and a clearer U.S. framework would likely help Circle’s institutional case. Large financial firms tend to like two things: liquidity and legal clarity. Stablecoins can deliver both, provided regulators stop treating every new payment primitive like a threat to civilization.

But regulation is not a substitute for product-market fit. It can remove friction. It cannot create demand out of thin air. Circle still has to prove that CPN, Arc, and the Agent Stack can generate durable fee revenue at meaningful scale. That is the longer-term test.

The bigger strategic question: is Circle building the future or repackaging a yield business?

Circle’s pivot makes sense. Frankly, it had to happen. A stablecoin issuer that depends mainly on reserve yield is living with a built-in ceiling. The smarter path is to own more of the transaction stack, create recurring network usage, and build products that work whether rates are 5% or 2%.

But there is a hard-nosed counterpoint worth keeping in view. Circle’s new strategy could deepen dependence on its own corporate rails even as it reduces dependence on the bond market. That may improve economics, but it also pushes the ecosystem toward a more centralized model. For a company in crypto, that tension matters. Efficiency is nice. Control is profitable. Decentralization, meanwhile, keeps getting invited to the meeting and then told to wait outside.

The real question is whether Circle can convert adoption into actual fee-bearing usage without sacrificing too much margin along the way. More activity is good. Better profit quality is good. But if the company keeps adding layers of infrastructure without meaningfully diversifying revenue, then the market may eventually decide it has not escaped the reserve-income trap — it has simply decorated it better.

If Circle pulls this off, USDC could become more than the most useful dollar proxy in crypto. It could evolve into the settlement asset for institutional payments, on-chain FX, cross-border transfers, and maybe even autonomous AI commerce. If it doesn’t, then the market will remember a company that talked a strong game about infrastructure while the reserve income was still doing most of the heavy lifting.

Key questions and takeaways

• What is Circle trying to become beyond USDC?

  Circle is trying to evolve from a reserve-yield-dependent stablecoin issuer into a broader digital-asset infrastructure company with payments, settlement, blockchain, and AI-payment products.

• Why does reserve interest income matter so much?

  Because it still makes up about 94% of revenue. That model is highly exposed to interest rates, which means Circle’s earnings can swing when monetary policy changes.

• What does Circle Payments Network do?

  CPN is Circle’s payments and settlement network for financial institutions. It helps Circle keep more USDC activity on its own rails and collect more of the economics from transfers and settlement.

• Why is Arc important?

  Arc is Circle’s planned Layer 1 blockchain for cross-border payments, settlement, and on-chain FX. If it gains real adoption, it could become a meaningful fee generator.

• Is Hyperliquid important to Circle?

  Yes. Hyperliquid is a major distribution channel for USDC and could drive a large share of circulation if its growth continues.

• Is Circle’s AI payments stack a real business yet?

  Not yet. It is a forward-looking bet on autonomous agents and micropayments, with most of the upside still ahead.

• What is the main risk in Circle’s strategy?

  The main risk is that Circle spends heavily to build infrastructure and still fails to convert adoption into durable fee revenue, while margin pressure eats into profits in the meantime.