How stablecoin issuers make money

The engine is interest on the reserve

The reserve is the pool of assets an issuer holds against the tokens in circulation: cash at banks, short US Treasury bills, overnight repurchase agreements and money-market funds. Reserve yield is the interest those assets earn. For a fiat-backed issuer it is almost the whole business.

The mechanism is simple. A client sends Circle 1 mn USD and receives 1 mn USDC. Circle puts most of the dollars into the Circle Reserve Fund, a government money-market fund managed by BlackRock, which buys Treasury bills. The client can hand the tokens back at any time and get the dollars back, so the reserve must stay safe and liquid. The interest it earns while it waits belongs to Circle, and the USDC holder receives none of it.

Issuers could charge fees for creating and redeeming tokens, but the large ones mostly do not charge verified institutional clients. A cheap, fast swap between tokens and dollars is what holds the peg, as the chapter on reserves and redemption explains, so fees would weaken the product that earns the interest.

This is why stablecoin issuers are often described as a spread business. The spread is the gap between what the reserve earns and what the issuer pays the holder, and for USDT and USDC the second number is zero.

Circle pays for its distribution

Circle is the only large issuer whose economics sit in audited public filings, because it listed on the NYSE in June 2025. Those filings show a business that earns a lot and keeps little.

Distribution costs are what Circle pays partners who put USDC in front of users, above all Coinbase, the largest US crypto exchange. From the first quarter of 2025 through the second quarter of 2026, those costs took between 58.6% and 61.8% of Circle's revenue in every published quarter. Circle calls what remains revenue less distribution costs (RLDC), its own share before salaries, technology and the costs of being a public company.

The reason for the size of that bill is a contract. When Circle and Coinbase dissolved their joint USDC consortium on 18 August 2023, they signed a collaboration agreement under which Coinbase receives all the reserve income on USDC held on Coinbase's platform and half of the remaining income on USDC held elsewhere. The agreement was renewed on 18 August 2026 through 2029 on the same structure.

Deep dive How the Coinbase formula works

The agreement, filed as an exhibit to Circle's 2025 annual report, defines two parts of Coinbase's payment. The first is proportional to USDC sitting in Coinbase's own custodial products, where Coinbase keeps the full reserve income. The second is 50% of a "residual payment base" covering USDC held anywhere else, such as self-custody wallets or other exchanges, after Circle's own allocation and other partners' fees. Some of the percentage tiers in the public exhibit are redacted. The practical effect is that USDC growth on Coinbase is worth much less to Circle than USDC growth anywhere else.

Circle disclosed the consequence itself. A rise of one percentage point in the reserve rate would add about 618 mn USD of reserve income in a year, and about 315 mn USD of distribution costs. Roughly half of every extra dollar of interest leaves the company before Circle pays its own staff.

Why accept such terms? Because Coinbase is the channel that reaches US retail and institutional users, and it co-founded USDC. The chains and distribution chapter makes the broader point: in 2026, getting a token into the right apps and exchanges is harder to copy than issuing one.

Tether keeps the spread

Tether, issuer of USDT, earns on the same kind of reserve and pays no Coinbase-sized partner. It is a private company, now based in El Salvador, and publishes quarterly attestations: an accountant's check that the reserve matched the tokens on one date, which is narrower than a full audit. The reserves chapter covers that difference.

The comparison is stark. Tether reported a net operating profit of about 1.5 bn USD for the second quarter of 2026. Measured against its reserve, that is about 3.2% a year. Circle's net income in the same quarter, measured the same way, was about 0.26% a year. Both reserves earn roughly the same gross rate on Treasury bills. The difference is what happens between gross and net.

Tether also holds assets that pay no interest: gold and bitcoin. They can make a good year look spectacular and a bad quarter look awful. Tether says it earned more than 13 bn USD in 2024, and a large part of that was a rise in the value of gold and bitcoin rather than interest.

In short, Tether's higher margin per reserve dollar comes from two absences: no distribution contract like Circle's, and none of the costs of a listed company. It is not a better return on Treasury bills. The price is less transparency and a buffer that moves with gold and bitcoin.

Why interest rates decide the business

If income is interest on the reserve, then income is the reserve times the interest rate. Nothing else in the model moves the top line as much. The US federal funds target range was 3.50% to 3.75% after the Federal Reserve held rates on 17 June 2026.

The table below is simple arithmetic, not a forecast. It applies round rates to USDT and USDC supply on 4 September 2026 and takes Circle's share after distributors at about 40%, inside the range Circle has given investors. It leaves out Tether's gold and bitcoin and Circle's smaller revenue lines.

Author's calculation on DefiLlama supply of 183.3 bn USD (USDT) and 74.5 bn USD (USDC), 4 September 2026.

The breakeven question follows. Circle's adjusted operating expenses were 144 mn USD in the fourth quarter of 2025, which puts a normal year at roughly 600 to 700 mn USD. At a 1% rate, Circle's share after distributors would fall well below that. Tether does not publish its operating costs, so its breakeven is unknown, but a small private team and no distribution contract put it much lower.

The effect is already visible. Circle's revenue peaked at 739.8 mn USD in the third quarter of 2025 and was 701.3 mn USD in the second quarter of 2026, while USDC supply was roughly stable. The drop came from lower rates on the reserve, not from fewer tokens.

Paying the holder anyway

US law blocks the obvious fix. The GENIUS Act, signed on 18 July 2025, bars issuers of payment stablecoins from paying holders interest or yield. The regulation chapter covers the law. The market's response was to keep the payment token plain and pay the return somewhere else.

A yield-bearing stablecoin is a token that passes some return to its holder, usually as a separate wrapped version of a plain stablecoin. Four structures were live in September 2026:

1. A distributor pays a reward. PayPal funds rewards on PYUSD itself, as a loyalty program, while Paxos issues the token and holds the reserve. Coinbase offers up to 3.50% a year on USDC to its paying members, funded from its share of Circle's reserve income.
2. A protocol sets a savings rate. Sky pays the Sky Savings Rate to holders of sUSDS, a staked version of its USDS. The rate was about 3.75% in the second quarter of 2026, down from above 8% in 2024, funded by crypto loans, tokenized real-world assets and USDC in its reserve.
3. A hedge passes through a market rate. Ethena holds crypto and shorts the same amount in perpetual futures, and passes the income to holders of sUSDe. The funding rate is the fee traders on one side of a perpetual contract pay the other, and it is the source of that income. That return swung from about 56% in March 2024 to about 4% in August 2024.
4. The product is a fund, not a stablecoin. Tokenized Treasury funds such as BlackRock's BUIDL pay their interest openly, because they are securities, not payment stablecoins.

Demand followed the return. In 2026, as rates and funding income fell, USDe supply shrank by about a third and PYUSD by about a sixth, while USDT and USDC, which pay nothing, barely moved. The market snapshot chapter shows the full picture.

Deep dive Which of these the law touches

The GENIUS Act's ban applies to issuers of payment stablecoins. PayPal's reward is paid by a distributor, not the issuer, and resembles card loyalty programs. USDe is not a payment stablecoin under the Act's definition, so the ban does not reach it at all. Sky argues that it has no issuing company in the ordinary sense, since parameters are set by holders of its governance token, and that argument has not been tested by a US regulator. The Office of the Comptroller of the Currency proposed in February 2026 that some affiliate yield arrangements fall inside the ban. No final rule on that point was in force in September 2026.

How new entrants try to win

A newcomer faces an incumbent with years of trust and a distribution network it has already paid for. Each new model answers the same problem: who gets the reserve income, and what do they do for it.

The first answer is to own distribution. PayPal did not need a Coinbase because it already had hundreds of millions of accounts; it rents issuance from Paxos and keeps the customer relationship. The second answer is to share the reserve income widely. Paxos's USDG, launched in November 2024 with the Global Dollar Network, shares reserve income with member firms such as Kraken and Robinhood according to how much USDG each of them brings in. Circle keeps the income and pays one large partner; USDG pays many small ones.

The third answer is to let a company issue its own token. Bridge, which Stripe bought for about 1.1 bn USD, offers what it calls Open Issuance: a business launches its own branded stablecoin and, Bridge says, keeps the reserve yield itself. The fourth answer is to lean on a single powerful venue. World Liberty Financial's USD1 was chosen as the settlement token for a 2 bn USD investment into Binance, and Binance became its main trading venue.

Circle's own answer is to earn something besides interest. It is building Arc, its own blockchain, and the Circle Payments Network, and in 2026 it raised its guidance for non-interest revenue to between 310 and 330 mn USD, mostly tied to Arc. That is still small next to reserve income of more than 2.5 bn USD a year.

What to watch next

• Federal Reserve rate cuts: each quarter-point cut removes about 0.19 bn USD a year from Circle's gross reserve income at current supply, so the path of rates is the main driver of issuer earnings.
• Circle's non-interest revenue: if Arc and the payments network grow into a large line, Circle becomes less of a rate bet. If they stay small, its margin keeps thinning as rates fall.
• The final US rule on affiliate yield: a strict rule would squeeze Coinbase's USDC rewards and PayPal's PYUSD rewards, and with them a reason to hold those tokens.
• USDG and other revenue-sharing tokens: steady growth would show that paying many distributors beats paying one.
• Tether's first full audit: KPMG began one in 2026, with no completion date set by mid-year. A clean result would narrow the gap between Tether's economics and its reputation.