This post is not really about the bill, though. It is about the thing the bill was carrying and implications for Stablecoin

As expected, the Digital Asset Market Clarity Act failed to move through the Senate.

Yesterday, the Senate cloture motion on H.R. 3633 fell 49–50 — eleven votes short of the sixty required to advance, and one shy of even a bare majority. Multiple Republicans voted no. Senator Chris Coons did not vote. Senator Cynthia Lummis made the closing pitch on the floor: “Do not let this day be the day we handed our future to someone else because we were too afraid to finish what we started” and could not find the votes. Bitcoin slid from near $80,000; Coinbase closed down almost 9%.

Nobody who was reading the tape should be surprised. I have been writing this outcome down in stages since February. In Stablecoin Rewards’ Last Hope – Clarity Act I put the odds at roughly 70% and named the two things that would sink it: the rewards fight and the calendar. In The CLARITY Act Is Locked I flagged the remaining hurdles: Senate Agriculture jurisdiction, sixty floor votes, House reconciliation, …etc. and wrote that if the floor count fell short the bill slides to the next Congress. Prediction markets took until August to agree, collapsing from 82% to under 20%, and sat at 14% on the morning of the vote itself. Market structure legislation is not a 2026 story and it may not be a 2027 story either.

What CLARITY Actually Was — The Short Version

Two jobs, and only two that mattered.

Job one: market structure. Define what a digital asset is in law. Draw the line between a “digital commodity” and a “digital asset security,” and set functional criteria — the “ancillary asset” provisions — for sorting tokens into one bucket or the other. This is the definitional gap that has driven a decade of regulation-by-enforcement.

Job two: the regulator. Assign jurisdiction. The House-passed text tilted toward the CFTC, granting it new authority as watchdog of crypto spot markets. The Senate rewrite broadened SEC authority over ancillary assets, which is precisely why Coinbase walked out of the January markup. The Tillis–Alsobrooks amendment pulled the balance back toward the CFTC. Autonomous Research read the amendment as balancing the two agencies “in favour of CFTC” (Rahul Jindal, Autonomous Research, “Fintech Flyby: Market Structure Update,” February 2026).

That is the whole architecture. Everything else — DeFi protocol treatment, the tokenised equities language in Section 505, developer safe harbours, the insolvency safe harbour, the ethics provisions on officials’ crypto holdings — was ballast bolted onto those two beams. And it was the ballast that sank it: the ethics provisions and the stablecoin rewards language were the two unresolved rifts going into the vote.

With no statute, the industry now depends entirely on agency discretion. The SEC has proposed Regulation Crypto Assets. CFTC Chair Selig has instructed staff to draft a market-structure regime under existing Commodity Exchange Act authority. Both are real. Neither is durable — as Enso’s Connor Howe put it after the vote, “The next chair can rewrite an agency rule without a single vote in the Senate.” Even SEC Chair Atkins has conceded the new rules won’t hold up without a law underneath them.

That is the headline story, and every crypto outlet is running it. It is not the important story for payments.

What the Crypto Lobby Was Actually Buying: A Yield Loophole

Here is the part that matters, and it is worth being blunt about it because the industry has spent eighteen months not saying it out loud. The single most commercially consequential provision in a 600-page market structure bill was a carve-out that would have let platforms and third parties pay something that walks and talks like interest on stablecoin balances.

The mechanics of the loophole are simple. The GENIUS Act, signed July 2025, prohibits issuers from paying interest or yield to holders. Section 4(a)(11) is written narrowly and the operative word is “solely”: no permitted payment stablecoin issuer shall pay interest or yield to a holder solely in connection with the holding, use or retention of the stablecoin. It says nothing whatsoever about exchanges, wallets, or affiliates.

So the money took a detour. Circle earns yield on USDC reserves — Treasury bills at 4.5%-plus. Circle keeps a slice and passes most of it to distributors. Coinbase captures roughly half of Circle’s reserve income; Circle’s distribution payments to Coinbase hit $364 million in Q4 2025 alone, against $151 million in Q1 2023 (Kenneth Suchoski, Autonomous Research, “Clouded Clarity,” March 2026). Coinbase then paid users roughly 4% APY on USDC balances. Stablecoin-related revenue was about $1.3 billion for Coinbase in 2025 (~ a fifth of the entire top line).

Three-party structure. Same economics as a savings account. No issuer ever paid a cent of interest to a holder. Legally clean under GENIUS as drafted. The lobby’s objective in CLARITY was to convert that structural accident into statutory permission. And the text moved in exactly that direction and then partially back:

• The House-passed version (July 2025, 294–134) cross-referenced GENIUS for stablecoin treatment and said nothing about third-party rewards — preserving the loophole by silence. Coinbase supported it.
• The Senate draft introduced Section 404, extending the prohibition from issuers to the whole intermediary ecosystem: “No digital asset service provider shall pay the holder of any payment stablecoin any form of interest or yield (whether in cash, tokens, or other consideration) solely in connection with the holding of such payment stablecoin.”
• The Alsobrooks Compromise then reopened the door at 404(b)(2) with a carve-out for “activity-based rewards”: payments tied to an affirmative action: executing a trade, buying goods, staking, providing liquidity, posting collateral.
• The final May text banned anything “economically or functionally equivalent to the payment of interest or yield on an interest-bearing bank deposit” while permitting rewards tied to “bona fide activities or bona fide transactions.“

When the March draft of that language leaked, Circle fell about 20% intraday and Coinbase about 10%. That is the market pricing a legislative sentence. It tells you exactly what the bill was worth to whom.

The American Bankers Association called the activity-based rewards provision “a significant loophole” that would let exchanges offer “interest-like incentives” through marginally different legal structures. America’s Credit Unions cited Treasury estimates of up to $6.6 trillion in potential deposit migration if stablecoins paid competitive returns. That number was always alarmist — but the direction was right, and it explains the intensity. Hilbert Group’s Barnali Biswal noted after the vote that “major bank trade groups were lobbying against the stablecoin yield language right up to the vote.”

They were lobbying right up to the vote because the vote was the last exit. They didn’t need to win the argument. They needed the clock to run out.

The Terms of GENIUS Will Hold — And the Fight Moves to Rulemaking

This is the core of it. No CLARITY means no statutory yield carve-out. GENIUS stands unamended, and GENIUS says no yield. What changes now is not the policy. It is the venue. The question stops being what will Congress permit and becomes what will the OCC enforce. And the OCC has already told us, at length, in writing.

The February 2026 proposal — 376 pages, comments closed 1 May — mirrors the statutory prohibition and then goes considerably further. I covered it in detail in No More Stablecoin “Rewards”, and the central mechanism is a rebuttable presumption of evasion:

If an issuer has an arrangement to pay interest or yield to an affiliate or related third party, and that party in turn has an arrangement to pay interest or yield — in cash, tokens, or other consideration — to a stablecoin holder solely in connection with holding or using that stablecoin, the OCC will presume the prohibition has been violated.

The burden of proof inverts. The issuer must come forward with written materials and demonstrate to the OCC’s satisfaction that the arrangement is neither prohibited interest nor an attempt to evade the prohibition. Sullivan & Cromwell noted the presumption is “not specified in the GENIUS Act” ; it is regulatory construction, built specifically to reach the three-party structure the statute missed. The definition of “related third party” expressly covers anyone paying yield to holders as a service on the issuer’s behalf, and anyone in a white-label relationship.

The carve-outs are narrow and deliberate: merchants may independently discount for stablecoin payment, and issuers may share white-label profits with non-affiliate partners provided nothing reaches the end holder as yield. That is it.

The enforcement architecture is not theoretical either — cease-and-desist orders, civil money penalties, revocation of permitted-issuer status, and a proposed extension of Prompt Corrective Action under 12 CFR Part 6 to stablecoin issuers. This is national-bank supervision machinery, pointed at firms that have mostly never been examined.

Comptroller Jonathan Gould told the Wyoming Blockchain Symposium in August: “We are very intent on moving quickly and getting a final rule out by November so that we will be able to start processing applications within the new year.” The GENIUS Act takes effect on the earlier of 18 January 2027 or 120 days after final implementing rules. Treasury’s own proposal on issuance, offer and sale published 18 August with comments open to 19 October. The FDIC’s proposed §350.3(b)(4) mirrors the prohibition for its supervised issuers.

Law to enforceable regulation, in roughly ninety days. No yield.

Where I Could Be Wrong

I have been early on this call, so let me be honest about the seams in it.

Rewards do not switch off on 16 September. The GENIUS prohibition binds issuers. The OCC’s reach into distributors runs through an issuer arrangement. A platform paying rewards from its own general revenue, with no yield-passing agreement from an issuer, is a harder case to presume against. Expect restructuring, subscription bundling. Coinbase already moved USDC rewards behind the Coinbase One paid tier. The direction is set; the last mile will be argued.

The market routes around fences. Simon Taylor’s objection is the strongest one against my position: “Stablecoins are programmable. A user could hold tokenized money market funds, automatically swapping into stablecoins at the point of transaction. Build the wrong regulatory fence, and the market routes around it instantly” (Simon Taylor, Fintech Brainfood, “The Stablecoin Opportunity That Banks Are Missing,” October 2025). He is right that the sweep model where idle balances are swept into a tokenized MMF like BUIDL at 3.5–5%, atomically redeemed at the moment of spend — satisfies every regulator and delivers the yield anyway. He is also right that USDe, offering 10%-plus through a delta-neutral hedge and entirely outside GENIUS compliance, is a live demonstration that suppressed compliant yield can push demand toward less compliant yield.

And his Durbin analogy deserves respect. In 2011 a community-bank carve-out accidentally created Chime, Cash App and a generation of billion-dollar fintechs. Taylor’s argument is that the yield loophole is structurally identical and would have bootstrapped an entire generation of on-chain consumer finance. That is a serious claim, and the honest answer is that we will now never run the experiment.

But note what the sweep model does to the payments thesis, which is the point I keep returning to. It does not make stablecoins a better consumer payment instrument. It completes their transformation into a treasury and savings product with a payment rail bolted on. The yield lives in a regulated security. The stablecoin stays a pure settlement token. Which is the thesis, arriving by a different road.

Suchoski frames the residual uncertainty better than I can: “If yield on USDC is curtailed, we would expect USDC supply and revenue to come under pressure as there’s likely some portion of current USDC demand driven by yield rather than utility. There are nuances as we don’t know what amount of activity is driven purely by demand for yield.” Nobody can cleanly separate yield-driven from utility-driven balances. I think the split favours utility at the institutional end and yield at the retail end. That is a judgement, not a measurement.

Implications for Consumers: Money at Rest Stays in the Account

This is the consequence that actually reaches households, and it is the least discussed. Strip out the yield and ask what remains as a reason for a consumer to hold a stablecoin balance rather than leave the money in a bank account. The honest inventory:

• Activity-based rewards. Which already exist, and are called credit card points. The average American earns 1.5–2% cash back on every dollar spent, with zero onboarding, zero gas fees, and no wallet to manage. Stablecoin rewards can in theory match card cash-back. They cannot structurally exceed it, and they are funded by platform balance sheets rather than a $50 billion interchange pool.
• Speed and finality. Not consumer value propositions. They never have been. A consumer does not experience settlement.
• Cost. Not at the consumer layer. I paid $0.53 in Ethereum gas fees on a $2.50 Stripe transaction, a 21% effective fee and wrote it up in Stablecoins Are Not Free. Consumers do not pay to use cards. They are paid to use cards.
• Access for the underbanked. Real, and the strongest remaining argument. Also a small share of the addressable market, and largely solved domestically by neobanks and prepaid.
• Cross-border remittance. Real, material, and genuinely better. Also a use case, not a wallet.

Theme: remove the yield and you remove the only reason for a consumer to hold a stablecoin balance rather than spend one.

The BIS found that adjusted stablecoin transaction values are roughly 1% of reported volumes, and that only 0.4% to 0.9% of that adjusted activity is under $250 (BIS Papers No. 170, May 2026). Consumer stablecoin payments are, empirically, a rounding error inside a rounding error. The yield loophole was not a bonus feature on a thriving consumer product. It was the product.

So the practical consumer outcome of 15 September is this: money at rest stays in the account. Balances that were parked on exchanges earning 4% will drift back toward banks, brokerages and money market funds, where the yield is legal, insured or regulated, and where nobody has to hold a seed phrase. Stablecoins keep the jobs they are actually good at — moving value, crossing borders, settling — and lose the job they were quietly borrowing from banking.

Implications for Stablecoin Specialists: The Business Model Just Got Narrower

For the firms whose income statements depend on this, the consequences are sharper.

Circle. The reserve-income model survives intact — Circle keeps earning on T-bills. What narrows is distribution. If the OCC’s presumption bites, the revenue share that buys shelf space at Coinbase becomes a compliance exposure rather than a growth lever. Suchoski’s framing stands: some portion of USDC demand is yield-driven and comes under pressure. Circle’s answer is already visible, and it is instructive — it took a national trust charter.

Coinbase. Roughly $1.3 billion of stablecoin revenue and about a fifth of the top line sit against a rule whose final text lands in November. The mitigation has already started: USDC rewards moved behind the Coinbase One subscription, which reframes yield as a bundled benefit of a paid service rather than a payment for holding. Whether that survives “economically or functionally equivalent” review is exactly the question the OCC left itself room to answer.

Everyone below them. The smaller platforms that used high-yield dollar balances as a customer-acquisition channel lose their acquisition channel. There is no cheaper substitute. CNBC noted in May that the compromise would “pressure smaller crypto platforms that have leaned heavily on high-yield deposit products to attract users” — that pressure now arrives via rulemaking instead, with less notice and no negotiated carve-out.

The strategic response is already in motion, and it tells you who won. If you cannot pay yield as a crypto firm, you become a bank — because banks are permitted to pay interest. Circle’s First National Digital Currency Bank received OCC approval this year, alongside Ripple National Trust Bank, Paxos, BitGo and Fidelity Digital Assets. Gould’s eightfold increase in digital asset chartering activity is not a curiosity; it is the entire industry migrating toward the one legal structure where paying for deposits is a feature rather than a violation.

Stablecoin issuers are becoming licensed banks.

Chalk Up 2026 as a Win for the Banks

They did not win a policy argument in Congress ( no bill passed). They won by successfully defending the perimeter of deposit-taking, and they won it twice over.

First, they kept yield inside the regulated deposit franchise. The single most valuable thing a bank owns is the right to pay for money and lend it out. GENIUS left that right untouched for banks and denied it to stablecoin issuers. CLARITY was the one vehicle that could have extended a functional equivalent to non-banks at national scale. It failed. The deposit franchise is intact, and the $6.6 trillion migration scenario is off the table for this Congress.

Second, they are now the ones issuing. This is the part that makes 2026 a structural win rather than a defensive hold:

• On 1 September, twenty-one institutions — Bank of America, Citigroup, Goldman Sachs, Wells Fargo, UBS, Deutsche Bank and Fidelity among them — committed to forming a jointly owned company to issue a USD stablecoin, targeting launch in the first half of 2027.
• JPMorgan, Citi and The Clearing House continue building tokenised deposits for 24/7 interbank transfer, which I covered in JPMorgan, Citi and TCH: Tokenized Deposits ON Chain.
• Visa built the Visa Stablecoin Platform as bank-first infrastructure — because, as I wrote then, banks trust banks, not Circle. 63% of corporates want stablecoin access through their existing bank rather than a new provider, with 54% intending to be live within six to twelve months (Simon Taylor citing EY Parthenon, Fintech Brainfood, October 2025).

So the settlement position for the year: stablecoin issuers are chartering as banks, banks are issuing stablecoins, and the yield stays in the deposit system. Every one of those three moves runs in the banks’ favour.

Taylor (a friend who is generally the sharpest critic of bank complacency) puts the underlying dynamic best: “when banks choose to innovate, they often win. Card networks, SWIFT, CLS, even Zelle, all dominant. All were built by banks willing to embrace new infrastructure.” That is what 2026 looks like in retrospect. The banks stopped resisting the technology and took ownership of it, while the legislative route to competing with them on deposits quietly expired.

The one caveat worth holding: a win on the perimeter is not a win on the product. Twenty-one banks forming a committee to issue a stablecoin in 2027 is a governance achievement, not yet a payments one. Consortium infrastructure has a long history of arriving late and under-featured. The deposit franchise is safe; whether the banks build something anyone wants to use is a different question, and one I have been sceptical about since Message to Bank CEOs as Stablecoins Take Hold.

My Consistent Position: Stablecoins are a settlement innovation

I have written this repeatedly, most directly in Stablecoins Are Not Free and again in the Visa Stablecoin Platform piece: the card is an interface, not a rail. Stablecoins are a rail. Rails do not need consumers to love them. They need throughput, finality and governance.

Every genuinely successful stablecoin use case is a settlement use case. Cross-border payouts. Treasury funding. Merchant settlement : Visa is running a $7 billion settlement run rate across 160+ programs. Exchange collateral. On-ramps and off-ramps. Correspondent banking replacement. B2B flows, where the cost of the incumbent rail is genuinely painful and the counterparties are institutions with compliance teams, not households with phones.

And the consumer story was always downstream of yield. That is why the yield fight was the real fight, and why the crypto lobby spent its political capital on a paragraph in Section 404 rather than on the jurisdictional architecture that the bill was nominally about. They understood the stakes precisely. They lost on the calendar.

So here is where 15 September leaves us. No market structure law. Agency rulemaking as the operative regime, finalizing in November, effective January. No yield for holders. Banks holding the deposit franchise and issuing the tokens. Stablecoin issuers becoming banks to get inside the perimeter.

And stablecoins settling into the job they were always good at.

Settlement has been the core of banking for as long as there has been banking ( Settlement: The Core of Banking). The technology changed. The function did not. Stablecoins are the newest instrument for performing the oldest job in the industry, and the removal of yield does not diminish that. It clarifies it.

Which is a better outcome than the CLARITY Act managed.