Stablecoins Make Billions, Just Not for Blockchains
Published on
Stablecoins are worth nearly $300 billion and sit on the other side of most onchain activity. Yet the money they generate barely makes it back to the blockchains that move them: in Q3 2026, Tether and Circle generated roughly $2.1 billion, compared with $285 million in fees for the thirteen blockchains we tracked. Here's a look at a business model where only a handful of players split the revenue.
Become Premium
Unlock all our research and get the right insights, at the right time.
Introduction
In less than seven years, stablecoins have gone from a simple tool for crypto trading to the backbone of all of onchain finance. Their market cap has grown from $4 billion in January 2020 to $298 billion today according to DefiLlama, after peaking at $310 billion in late April. To put that in perspective, that's more than three times the TVL of all of DeFi.
Beyond their size, they're really the foundation of onchain finance. Stablecoins serve as collateral on perps, liquidity on DEXs, capital on lending protocols and, more broadly, a store of value for users. Today, they account for nearly 28% of assets deposited on Aave, 37% on Morpho and 84% on Hyperliquid. Simply put, a large share of onchain activity couldn't exist without them.
Of course, the relationship goes both ways. It may sound obvious, but without blockchains, stablecoins would just be a digital dollar with little to offer. Sure, they'd still exist on CEXs, but there would be no DeFi protocols to give them a use, no 24/7 transfers to keep them moving, and so on. Blockchains are what created the market where issuers make their money today.
So you'd expect blockchains to get paid for the value they bring, and for revenue to be shared between those who issue stablecoins and those who keep them alive. That's not the case. Not even close.
In Q3 2026, Tether and Circle generated roughly $2.1 billion, compared with $285 million in fees for the thirteen blockchains we tracked. In this analysis, we compare what issuers and blockchains actually earn, take a closer look at Tron, then at the players who really capture this rent. Let's dive in.
Become Premium
Unlock all our research and get the right insights, at the right time.
Two Business Models That Have Nothing in Common
Issuers Earn on the Stock
To understand this gap, you first need to look at how each side makes money. On the issuer side, the model is remarkably simple. When you buy 1,000 USDC, Circle holds $1,000 in its coffers, which it invests in US Treasury bills, repos and cash. You hold a token that earns you nothing, but Circle doesn't just sit on those dollars: it collects interest on your cash.
In fact, that's the bulk of its business. In Q2 2026, Circle's reserve income made up 95% of its $701.3 million in total revenue. In other words, its business really boils down to one thing: supply multiplied by the yield on reserves.
You'll notice that usage doesn't show up anywhere in that equation (at least not directly, since we can all agree that the more USDC gets integrated, the more supply grows). The numbers back this up: between Q2 2025 and Q2 2026, USDC's onchain volume jumped 151% to $14.8 trillion, and its average supply grew 25%. Yet according to Circle's quarterly report, its reserve income rose just 5%, to $667.7 million.
The reason is simply that over the same period, reserve yield fell 66 basis points, from about 4.15% to 3.49%, as the Fed cut rates. Tether follows the exact same logic: its average supply grew 10.6% year over year, but its revenue as estimated by DefiLlama stayed stuck around $1.5 billion per quarter. It's that simple: whether USDT sits idle in a wallet or changes hands a thousand times a day, Tether earns the same.
Blockchains Earn on the Flow
For blockchains, it's the opposite. A blockchain earns nothing on the dollars it hosts. Whether it's $1 million or $1 billion, it only gets paid on the total number of transactions, through the fees users pay. That means its revenue depends solely on the activity it processes (and yes, once again, more stablecoins means more activity, but the point stands: TVL itself doesn't get paid).
Here's a pretty funny observation, by the way. If blockchains were paid for the stablecoins they host, Ethereum would be the most profitable blockchain by far, with $142 billion in supply today versus just $94 billion for Tron, roughly $15 billion for Solana and $13 billion for BNB Chain.
That's not the case at all. Take Ethereum and Solana. According to DefiLlama, Ethereum hosted an average of $146 billion in stablecoins in Q3 2026, more than half the supply across the major blockchains, but collected only $34.6 million in fees over the period. Solana, with less than $16 billion in stablecoins, collected $62.9 million. In other words, about 10x fewer stablecoins, and yet almost twice the revenue.
Some blockchains go even further, at least for a while. Robinhood Chain, which hosts only about $800 million in stablecoins, collected $46.2 million in fees over the quarter, more than Ethereum. That figure is mostly explained by the memecoin frenzy in September, when its fees at times topped $5 million a day (they've since fallen back to around $30,000 a day).
Now, these fees cover all of a blockchain's activity, not just stablecoin transfers. But since most of that activity (swaps, perps, lending, etc.) runs on USDT or USDC, this is ultimately the most favorable comparison possible for blockchains. And even then, a large share of those fees doesn't even go to them: Solana, for example, kept only $6.7 million of its $62.9 million in Q3, with the rest going to validators.
A Gap That Keeps Widening
To measure this gap, we estimated what Tether and Circle earn on the stablecoins hosted by each blockchain by applying their yield to that supply, then compared it with that blockchain's fees.
Take Ethereum. In Q1 2024, it still collected $1.17 billion in fees, versus roughly $730 million in revenue for issuers on the stablecoins it hosts. At that point, the blockchain was earning more than the issuers.
Two years later, the situation has completely flipped. In Q3 2026, Ethereum collects only $34.6 million, while issuers earn roughly $986 million. In other words, for every $1 Ethereum collects, issuers now earn $28.
To be fair, this is largely the result of Ethereum's own choices. With the Dencun upgrade in March 2024, followed by successive gas limit increases, the team deliberately lowered transaction costs on the Layer 1, and much of the activity moved to Layer 2s. It was a conscious choice: make the blockchain cheaper to make it more usable.
But at the end of the day, every blockchain makes its own technical and economic choices, and some fare better than others. Solana bets on massive transaction volume at very low cost, Tron on USDT transfers, BNB Chain on its ecosystem's trading activity. Results vary, but the trend is the same everywhere: competition between blockchains pushes the cost of a transaction toward zero. None of them wants to be the most expensive, and all of them try to attract activity by cutting fees.
That's the heart of the problem. A blockchain's revenue depends on a price that competition keeps driving down. An issuer's revenue depends on an interest rate set by the Fed, which no competitor has any control over. As long as that's the case, the conclusion won't change: the flow will always pay less than the stock.
Tron, the Textbook Case
A Blockchain Built Around USDT
You could argue that Ethereum is a special case, since stablecoins are only part of its activity. So let's look at the blockchain most dependent on stablecoins: Tron.
Unlike Ethereum or Solana, Tron is barely used for onchain finance. JustLend alone accounts for nearly 80% of deposits across Tron protocols, and that liquidity is itself highly concentrated: in 2024, Arkham identified a single wallet that had deposited 60,000 BTC into the protocol. But that's not what this article is about. In reality, Tron is used mainly for one thing: moving USDT.
USDT accounts for 98% of the $94.5 billion in stablecoins it hosts. That supply nearly doubled in under three years, from $48.8 billion in January 2024 to $92.5 billion in September 2026. In fact, Tron is well ahead of Ethereum in USDT and holds about 50% of all onchain USDT.
Credit where it's due, the strategy paid off. Very low fees applied to colossal volume brought Tron $577.6 million in 2024, then $579.3 million in 2025. In Q3 2026, Tron is even still the blockchain collecting the most fees among those we tracked, at $77.8 million, which it keeps in full since they correspond to burned TRX. According to Blockworks, nearly 90% of its revenue comes from USDT.
Ten Times Less Revenue Than Tether
What's pretty surprising is that even in this ideal case, the issuer earns far more. For every dollar of USDT hosted on Tron, Tether currently earns about 3.3 cents a year from its reserves. Tron, on the other hand, collects about 0.33 cents in fees. Across Q3 2026, that comes to roughly $745 million for Tether versus $77.8 million for Tron. Two years earlier, the gap was "only" 4.4x.
Why did it widen so much? Simply because Tron had to cut its prices. On August 29, 2025, the price of energy consumed by transactions dropped from 210 to 100 sun. Average daily fees fell 40% over the following three weeks, while the number of transactions grew only 3.4% between July and September.
In our view, this cut is hard to separate from competition. Less than a month later, on September 25, 2025, Plasma, a blockchain backed by the Tether ecosystem, launched with free USDT transfers. The goal was clearly to take on Tron by becoming the go-to blockchain for stablecoin transfers. Worth noting: despite zero fees, Plasma never really managed to dethrone Tron. Since its launch, USDT hosted on Tron has grown from $75.7 billion to $92.5 billion, while Plasma fell back from $6 billion to $1.3 billion. So Tron kept its users, but at the cost of a 40% drop in fees.
Above all, this is where you can see the limits of a flow-based model. When the only service you sell is transport, and a competitor can offer it for free, your price can only go one way.
How Is Revenue Really Shared?
Distributors Are the Winners
Make no mistake, issuers don't keep everything for themselves. They actually share a big chunk of their revenue, just not with blockchains: with those who control users and their balances. In H1 2026, Circle paid $819 million to its partners, or 59 cents for every dollar of revenue, for just 7 cents of net income.
The top beneficiary is Coinbase. Under their agreement, renewed through 2029, Coinbase keeps 100% of the revenue generated by USDC held on its platform, and 50% of the revenue from USDC held elsewhere. In Q2 2026, Circle's payments to Coinbase reached $324.6 million, nearly half of reserve income. Binance followed the same path: after an initial $60.25 million payment in 2024, Circle signed a five-year agreement with the exchange in September 2026, along with a $100 million equity investment.
Newcomers go even further. Paxos passes up to 100% of USDG's yield to the 150+ members of its Global Dollar Network, and kept only 47% of its revenue in Q3 2026, versus 100% through Q2 2025 according to DefiLlama. As for OUSD, launched on September 30 by Coinbase, Stripe, Visa, Mastercard and Shopify, who split its yield, it reached $835 million in barely a month.
And regulation only reinforces this logic. The GENIUS Act bans issuers from paying yield directly to stablecoin holders, but not the platforms that distribute them. The OCC confirmed this flexibility in its final rule of October 21, 2026. In other words, to attract and retain holders, an issuer has no choice but to go through those who already have them.
DeFi and Blockchains Walk Away Empty-Handed
This may be the most important part of it all. Onchain finance protocols, which give stablecoins most of their utility, aren't any better off. The stablecoins deposited in them, at least $16.85 billion, bring issuers roughly $550 million a year. On Aave alone, the $8.25 billion in deposited stablecoins generate about $270 million a year for Tether and Circle, but nothing for Aave.
As for blockchains, we're not aware of any recurring yield-sharing agreement with an issuer. Worse, the money actually flows the other way. The TON Foundation allocated 11 million TON in 2024 to reward USDT users, and Arbitrum is currently offering 100 million ARB in incentives to join Paxos's Global Dollar Network, with a vote scheduled from October 29 to November 12.
So it's blockchains paying to attract stablecoins, not the other way around. The reason, obviously, is that they're all fighting over the same pool of liquidity, and you have to be the most attractive to take market share from the competition. But as we saw above, even that isn't enough to generate revenue.
Can Blockchains Take Back Control?
Faced with this reality, several projects have tried to change the rules of the game. Some wanted to redistribute value differently, others launched their own stablecoin to capture reserve yield. The results are very uneven, and they say a lot about what actually works.
Berachain and Its Redistributive Model
Berachain is a particularly ambitious attempt. As we explained in our overview of the blockchain, the team's thesis was that Proof of Stake (PoS) isn't a model that fosters ecosystem cohesion. Validators capture all of a blockchain's revenue, even though protocols are the ones driving activity, and conversely, validators get no allocation when a protocol launches its token.
To fix this, Berachain created Proof of Liquidity (PoL). Validators distribute emissions of BGT, the governance token, to "reward vaults" where users deposit their liquidity, and protocols pay validators to direct those emissions toward their pools. The whole thing relied on a native stablecoin, HONEY. In other words, a system where validators, protocols and users all get paid for the value they bring. On paper, the model was solid.
But as with any product, it had to face reality on the ground: did users actually want it? The answer is no. Berachain's TVL went from nearly $3 billion in March 2025 to around $33 million today according to DefiLlama. Hosted stablecoins dropped from $1.09 billion to $58 million, HONEY from $638 million to $7.1 million, and the blockchain collected only about $6,000 in fees over all of Q3 2026. As we detailed six months after its launch, most of the capital had come for yield, and it left as soon as yields dropped.
Ethena and White-Label Stablecoins
The second approach is more pragmatic. Since September 2025, Ethena has offered blockchains and protocols the ability to launch their own white-label stablecoin, as we analyzed in our article on Ethena. The idea is that instead of integrating USDC or USDT and leaving the yield to Circle or Tether, a protocol or blockchain can issue its own dollar, backed by USDtb, USDe and USDC. Ethena handles issuance and reserve management, and the "client" keeps control of the yield.
It was clearly the most complete proposal to finally let onchain finance players plug the value leak from stablecoins. Several projects jumped in and adopted Ethena's white-label offering: MegaETH with USDm, whose yield funds MEGA buybacks, Jupiter with jupUSD on Solana, and Sui with suiUSDe. And ether.fi just announced its own.
But the results are still too modest. According to DefiLlama, USDm has fallen back to $16.1 million, down from $272 million in April, suiUSDe stands at $12.4 million and jupUSD at about $42 million. USDsui, launched by Sui with Bridge, has reached $78.5 million. By comparison, USD1 stands at $3.91 billion, USDG at $3.07 billion and PYUSD at $2.94 billion.
Still, it would be wrong to say the problem lies with the stablecoin itself. It's not just MegaETH's USDm that collapsed, but the entire ecosystem, whose stablecoin supply went from $661 million at the end of May to about $18 million today. Users didn't leave MegaETH because they preferred USDC to USDm, but simply because they no longer found enough reason to stay.
That's the real lesson from these attempts. It's not a problem of approach, but of market reality. Today, the vast majority of new blockchains and new protocols struggle to retain their users, with or without an in-house stablecoin. Those who tried to change the model are no exception; they simply met the same fate as most of the others.
Is Hyperliquid the Exception?
Only one player has managed to make a major issuer bend: Hyperliquid. In September 2025, with more than $5 billion in USDC sitting on its platform, Hyperliquid put the issuance of its native stablecoin, USDH, up for auction. The bids were spectacular, with Paxos and Ethena offering to redistribute 95% of the yield, and Agora and Frax 100%. Native Markets ultimately won, but USDH never topped $102 million.
Yet it was this bidding war that changed the game. In May 2026, Coinbase and Circle agreed to the AQAv2 mechanism. In short, USDH was dropped, and USDC now passes through about 90% of its reserve yield, net of costs, on the supply hosted by Hyperliquid. According to several media reports, the first payment reached $14.58 million for thirty days, or about 2.6% annualized on the $6.9 billion in USDC hosted. To put that in perspective, the Hyperliquid blockchain collected only $2.7 million in gas fees over all of Q3, a period three times as long.
What's interesting is that Hyperliquid didn't get this deal as a blockchain, but for exactly the same reason as Coinbase or Binance. The platform built infrastructure so performant, and liquidity so deep, that users leave their funds there. That's what's known as "sticky liquidity." Stablecoins also make up 84% of the assets deposited on Hyperliquid. In a way, Hyperliquid really does hold its users' balances, like an exchange, and chooses the reference stablecoin for its markets.
That's the leverage Circle and Coinbase couldn't ignore, and it's also what makes it hard to replicate. Aave, for example, holds billions in stablecoins, but doesn't choose which ones its users deposit.
When Issuers Build Their Own Blockchains
Finally, the last trend runs in the opposite direction. It's no longer blockchains trying to become issuers, but issuers and payments players launching their own blockchains dedicated to stablecoins. What they all have in common is near-free transfers.
On Tether's side, two projects have emerged. The first is Plasma, backed by Bitfinex, Founders Fund and the Tether ecosystem, which raised $373 million and launched its mainnet on September 25, 2025 with completely free USDT transfers. The start was impressive, with $6 billion in stablecoins by the end of September, but much of that money had come for XPL rewards. According to DefiLlama, Plasma now holds just $1.33 billion in stablecoins, a good chunk of it deposited on Aave. XPL has lost about 96% from its high.
Stable, the other blockchain tied to the Tether ecosystem, uses USDT directly to pay fees. Its first pre-deposit phase, capped at $825 million, was filled within minutes in October 2025, largely by a handful of wallets. Today, Stable hosts barely $19.3 million in stablecoins.
The other two projects are more recent. Tempo, incubated by Stripe and Paradigm and launched on March 18, 2026, has no native token. Fees, around a tenth of a cent, are paid directly in stablecoins. Its validators include Stripe, Visa and Zodia Custody (Standard Chartered), and it already hosts $613 million in stablecoins, mostly OUSD, for just $3,088 in fees in Q3. As for Arc, Circle's blockchain launched on September 16, 2026 with BlackRock, the DTCC and Visa among its validators, it uses USDC as gas and is already approaching $1 billion in USDC in six weeks.
These blockchains have one thing in common: they're not trying to make money on transfers. Plasma, still the largest with $1.33 billion in stablecoins, has collected only $268,916 in fees since launch, and Tempo barely $3,088 in Q3. Their backers make their money elsewhere: on reserves for Tether and Circle, and on payments for Stripe. For them, the blockchain is just a way to control the entire distribution chain, from issuance all the way to usage.
Even when issuers build the blockchain themselves, they don't try to make it profitable. As long as the rent sits in the reserves, transport has no reason to cost anything. That said, by controlling the entire distribution chain, they spare themselves a potential Hyperliquid-style showdown.
Conclusion
Ultimately, the numbers tell a pretty simple story. In Q3 2026, Tether and Circle generated more than seven times the fees of all the blockchains we tracked combined, even though those fees include all of their activity, not just stablecoins. In early 2024, the two were still roughly neck and neck.
The reason is that an issuer gets paid on the stock, at a rate set by the Fed, regardless of whether its dollars sit idle in a wallet or circulate. A blockchain gets paid on the flow, at a price that competition relentlessly pushes toward zero. Even Tron, the blockchain best positioned to benefit, had to cut its fees by 40% to keep its users.
And when the rent is shared, it goes neither to blockchains nor even to the onchain finance protocols that give stablecoins their utility in the first place. It goes to those who control users: Coinbase, Binance, Paxos's partners and, more recently, Hyperliquid. It's almost paradoxical. Blockchains created the market where issuers make their money today, yet they're the ones who benefit the least.
There's been no shortage of attempts to change this model, from Berachain to Ethena's white-label stablecoins. If they failed, it's not so much because the idea was bad as because users didn't show up. In the end, only those who truly control users and their balances manage to capture a share of the rent. Issuers have clearly gotten the message, since they're now building their own blockchains so they no longer depend on anyone.
The question is how long this model will hold. Falling rates mechanically shrink every issuer's rent, and therefore what's left to share. The OCC rules taking effect on January 18, 2027 will also show how far distributors can keep passing that yield on to their users. But one thing is certain: as long as the value sits in the reserves, it won't be blockchains getting rich off stablecoins.
Become Premium
Unlock all our research and get the right insights, at the right time.
