Here is the trap in reading Q2 as a regulatory retreat: the category most exposed to Washington's yield ban is the one that grew, and the one nobody is trying to protect is the one that collapsed. Two numbers inside the same quarter tell opposite stories, and the market has already picked which one it believes.
Yield-bearing stablecoin supply fell 15% in Q2 2026 — more than $3.5 billion — its first quarterly decline since Q3 2023, per CEX.IO's report. That headline reads like fear. It isn't.
Underneath it, two halves moved in opposite directions. Ethena's sUSDe lost 52% of its supply, shedding nearly $2 billion. Sky's sUSDS fell 16%.
Meanwhile Treasury-backed products expanded. Ondo's USDY rose more than 66%. Circle's USYC gained nearly 16%. BlackRock's BUIDL added 2%.
So the category didn't shrink. It sorted itself. Crypto-native yield imploded while reserve-backed yield grew, and the net was a decline only because the crypto-native cohort was so much larger going in.
The category didn't shrink. It sorted itself. Crypto-native yield imploded while reserve-backed yield grew, and the net was a decline only because the crypto-native cohort was so much larger going in.
Last week, this thread was about an open question: whether the CLARITY yield fight — the Tillis-Alsobrooks compromise, JPMorgan's shadow-banking campaign — would decide the fate of stablecoin yield. This week we have the first quarterly data showing capital rotating out of one kind of yield and into the other. The market started answering before Congress finished asking.
Start with the mechanism, because it explains everything downstream. sUSDe's "yield" was never reserve income. It came from the funding rate on perpetual futures — the premium longs pay shorts when trading is bullish and leverage is stacked.
Ethena earned that premium by holding spot crypto and shorting the perp, capturing the basis. When trading momentum is strong, the spread is fat and the yield is high. That's the entire engine.
But it runs in reverse just as fast. When momentum fades and leverage unwinds, funding compresses, the yield evaporates, and holders redeem. Supply doesn't drift lower — it retracts violently, because the thing anchoring it was never a deposit relationship. It was a bet on other people's leverage.
Q2 confirmed the timing. Talos flagged declining stablecoin supply alongside spot Bitcoin ETF outflows and slower Strategy purchases — three demand channels weakening together. Total stablecoin transaction counts fell by 530 million to 4.48 billion, reportedly the largest quarterly drop on record per CEX.IO.
When the trading tide went out, the yield tethered to it went with it. A 52% supply collapse in a single quarter is not a store of value repricing. It is a leverage product doing what leverage products do.
Now hold that against JPMorgan's argument. Jamie Dimon pledged to fight the CLARITY yield provision "down to the wire," and a JPMorgan executive called it "shadow banking by another name" — bank-like deposit products that could trigger mass withdrawals in a stress event, without the capital, insurance, or supervisory guardrails banks carry.
The run dynamic JPMorgan warns about is real. It just already happened — inside the crypto-native cohort, in Q2, with no stress event required. sUSDe ran on its own, on nothing more than fading trading momentum.
And here's the part that should reframe the whole fight: it ran toward Treasury reserves. The capital fleeing crypto-native yield didn't leave the yield category. It rotated into USDY, USYC, BUIDL — the products whose "yield" is short-dated Treasury income, the structure regulators worry about least.
This is the irony the data exposes. The Tillis-Alsobrooks compromise bans rewards offered "in a manner that is economically or functionally equivalent to the payment of interest or yield on an interest-bearing bank deposit." That language is engineered to stop deposit flight — the exchange-funded, promotional, deposit-mimicking yield the ABA reportedly says could put $6.6 trillion at risk.
But sUSDe's yield was never deposit-equivalent. It was a trading derivative wearing a stablecoin's clothes. The provision aimed at protecting bank deposits is pointed at a category that is structurally dying without any help from Congress — and the survivor of that die-off is the reserve-backed model, which behaves like a money market fund, not a shadow deposit.
So what does a yield ban actually accomplish here? It doesn't kill the threat. It accelerates the rotation already underway.
If CLARITY passes with hard prohibition language, or if the OCC and FDIC finalize anti-evasion rules broadly enough to reach promotional rewards, the crypto-native experiment loses its remaining regulatory oxygen — and the capital keeps flowing where it was already flowing. Into BlackRock, Ondo, and Circle's Treasury-backed products, which sit closest to the compliant structure the GENIUS Act's 93-day-Treasury reserve rules already prescribe.
That is the strategic content of the fight nobody is saying out loud. A yield ban would entrench the largest, most Treasury-native issuers by legislating away the only competitor category — and that category is self-liquidating anyway.
The macro tail makes this concrete. A BIS working paper reportedly found that a $3.5 billion five-day stablecoin inflow lowers 3-month T-bill yields by 2.5 to 3.5 basis points. Roughly $3.5 billion just exited the yield-bearing category in Q2 — the mirror image of that flow.
Stablecoins are already wired into the front end of the Treasury curve in measurable ways. When crypto-native supply retracts, that transmission runs backward: less reserve buying at the short end, small but real. The point is that the two halves of this category interact with the actual financial system through entirely different plumbing — one through perp funding rates, one through T-bill demand — and only the second is what the yield provision claims to be about.
Watch the timeline, because the delay is doing the work. Galaxy Research puts CLARITY's odds of enactment this year at roughly 50-50, possibly lower, with the stablecoin yield language among the unresolved disputes. The Senate wanted a July floor vote before the August recess; the bill is stuck in Banking Committee.
Every week of gridlock is a week the market keeps writing the answer Congress hasn't. The ABA's $6.6 trillion alarm and the White House CEA's $2.1 billion lending estimate have been arguing past each other on projections for eighteen months. Q2 just delivered the first real data — and it shows the yield most likely to survive is the one built on reserves, not on leverage.
None of this means the deposit-flight fight is imaginary. Exchange-funded rewards on USDC-style balances are a distinct threat, and the Tillis-Alsobrooks loophole the ABA is still contesting — permissible rewards calculated by "duration, balance and tenure" — keeps that door ajar. That fight is live and worth having.
But conflating it with crypto-native yield is a category error the Q2 data now makes hard to sustain. sUSDe didn't collapse because regulators threatened it. It collapsed because it was never a deposit substitute — it was a synthetic long on funding rates, and funding rates turned.
The durable question was always which yield model survives contact with a full market cycle. Q2 answered it in the cohort's own numbers: reserve-backed yield grew through the downturn, leverage-backed yield lost half its supply.
Congress is preparing to ban a threat the market is already retiring, in a way that would harden the position of the issuers best equipped to comply. The vote may still matter for the deposit-mimicking edge cases. For the crypto-native category, the verdict is already in — and it wasn't written in Washington.