Newsletter · · Ashutosh Agarwal

Clarity Act Senate Vote Looks Set to Fail as Banks Fight Stablecoin Yield - Stablecoins Eat Banking - Week of September 14, 2026

Stablecoins Eat Banking podcast synthesis for the week of September 8 to September 14, 2026. Coverage of the Clarity Act's Tuesday Senate cloture vote and its endangered stablecoin-yield ban (Section 404), banks going live on stablecoin and tokenized-deposit rails, Circle's $400 million Tazapay acquisition and ARC chain launch, Coinbase arming community banks, and an emerging neobank consolidation wave.

Stablecoins Eat Banking

Week of September 14, 2026: Clarity Act Senate Vote Looks Set to Fail as Banks Fight Stablecoin Yield

This is the week we've been circling on the calendar since July. Tomorrow, Tuesday September 15 at 2pm, the Senate holds the procedural vote that decides whether the Clarity Act, and with it the fate of Section 404, the clause that would ban stablecoins from paying you interest, moves forward this year. And the honest read from the people who count votes for a living is: it probably won't. Meanwhile, while Washington argues about whether banks should be allowed to use this technology, the banks just went ahead and used it, live, on real rails, moving real dollars. Two stories, one point: the incumbents stopped waiting for permission.

TL;DR

  • The Clarity Act vote is tomorrow, and the smart money says it fails. An analyst on Paul Barron's show pegged the odds of the September 15 procedural vote succeeding at 5%; prediction markets sit around 16%. Republicans hold 53 seats and need seven Democrats, but two Republicans (Hawley, Moran) are public "no" votes over the yield clause, a third (Tillis) is a "no" on ethics, McConnell is absent recovering from a fall, and zero Senate Democrats have publicly committed to yes (Paul Barron, Sep 8). The bank lobby escalated its fight against stablecoin yield the same week (Thinking Crypto, Sep 12).

  • Bank co-option stopped being a slide and became a live transaction. US Bank, the fifth-largest US commercial bank, ran its first live cross-border payment on its own USBDC stablecoin (on the Stellar blockchain); DBS and Citi settled a weekend dollar payment as tokenized deposits on SWIFT's new blockchain ledger, the second live SWIFT-ledger transfer after HSBC and Standard Chartered in August (Thinking Crypto, Sep 11; Thinking Crypto, Sep 8).

  • Circle went shopping and its own chain goes live the day after the vote. Circle bought cross-border payments firm Tazapay for $400 million (Tazapay does $25B+ a year, ~60% already in stablecoins), and its ARC blockchain launches on mainnet September 16. Circle's chief economist laid out why: banks and governments need settlement guarantees today's chains can't give them (DeFi Decoded, Sep 10; Thinking Crypto, Sep 9).

What's new

1. The vote is tomorrow, and the yield ban is the whole fight

Here's the mechanics, in plain English. Tuesday at 2pm the Senate holds a "cloture" vote, a procedural vote on whether to allow a vote. It needs 60 of 100 senators. Clear it, and the bill goes to debate, amendments, and a final floor vote. Miss it, and, in the words of the best-informed guest of the week, "it's done" for this attempt, though the bill can come back in the lame-duck session (Thinking Crypto, Sep 12).

The most credible voice was Ron Hammond, head of policy at trading firm Wintermute (insider, policy), who does this for a living. His update: the final ~600-page text just dropped with roughly 100 technical changes (small tweaks to anti-money-laundering and DeFi provisions), so the text isn't the problem. The problem is the whip count, and the banks. He was blunt that the bank lobby has "doubled down aggressively," going to senators who've never touched crypto with a scare story:

"All lending in rural Mississippi or, you know, pick a state is going to go away now because of this thing called stablecoins that you've never heard of… and there's a loophole, and you forgot about the loophole, and we're here to change it."

The tell for our thesis: the fight is entirely about stablecoin yield, Section 404. Hammond confirmed that the American Bankers Association, the Independent Community Bankers of America, and other bank trades came out against the Clarity Act on September 11, saying "major changes still need to be done on the yield issue", even though the industry already gave up ground in committee. In his words, the crypto side "made a compromise, we're adults, this is how it works. And the banks have just done the opposite and fought at every single turn" (Thinking Crypto, Sep 12).

Now the arithmetic, which is grim. An analyst on Paul Barron's show walked the count: "My probability that cloture succeeds on September 15th is 5 percent. That's almost no chance." Republicans hold 53 seats; if all 53 vote yes, cloture still needs seven Democrats. But Hawley and Moran are public 'no' votes over yield, Tillis is a 'no' without better ethics provisions, McConnell is absent recovering from a fall, and "the count of Senate Democrats who have publicly committed to yes on the floor is zero. Five percent is generous." Prediction-market odds were diving through the week, quoted around 16% (Paul Barron, Sep 8) [DIRECTIONAL, prediction-market-style odds]. Senator Lummis, throwing what one host called last-minute Hail Marys, warned that if Clarity doesn't pass this Congress, the next real window for market-structure law is 2030. And the structural obstacle from last week is confirmed: the House cancelled the final two weeks of its September schedule, so even a Senate win might not reach the President's desk this year.

There's a genuine two-sided view here, worth noting so we're not just doom-posting. The CoinDesk crew struck a more hopeful note, general optimism about getting to 60 if the White House ethics language comes through (CoinDesk, Sep 11), and Banking with Interest hosted the bank-side view of how the industry "clouded" the bill's future (Banking with Interest, Sep 8). But the vote-counters are bearish, and they're the ones I'd trust on a Tuesday.

Why it moves a number: if cloture fails, Section 404 doesn't get decided tomorrow at all. The yield ban that's meant to protect bank deposit margins stays un-passed, and the yield-bearing challengers keep scaling into the gap. As a respected trader framed it last week, that means the downside of a failed vote is largely priced, so the asymmetry now sits with a surprise pass, not a miss.

2. Co-option stopped being a forecast, the banks went live

Last week the incumbent-adoption story was a 21-bank press release about 2027. This week it's transactions that actually cleared.

Two landed. First, US Bank, the fifth-largest commercial bank in America, completed its first live cross-border payment using its own stablecoin, USBDC, on the Stellar blockchain. USBDC is a proprietary US-dollar-backed coin spanning US Bank's entities in North America and Europe; the pilot tested minting and redeeming the coin, plus freezing and clawing tokens back, and validated the bank's in-house digital-asset platform wired into its compliance, risk and finance systems (Thinking Crypto, Sep 11, Tony Edward, host, pundit, reporting the bank's disclosure).

Second, and arguably more important for the plumbing: DBS and Citi settled a weekend US-dollar payment as tokenized deposits on SWIFT's new blockchain ledger. That's the second confirmed live transaction on SWIFT's ledger, after HSBC and Standard Chartered ran the first cross-border transfer in August. SWIFT, the 53-year-old messaging network that underpins global banking, rolled out its blockchain ledger in July with 17 banks, and is racing to prove it can settle around the clock like the stablecoin rails threatening it. The host drew the distinction that matters for our beat: tokenized deposits are the banks' answer for moving millions and billions between institutions (wholesale, B2B), while stablecoins are for retail (Thinking Crypto, Sep 8, pundit).

Also this week: MoneyGram launched a USDC-powered Visa card in Colombia, the legacy remitter continuing to bolt itself onto the new rails (Thinking Crypto, Sep 11).

Why it moves the thesis: the co-option side is winning the infrastructure war in real time. The banks aren't studying stablecoins and tokenized deposits anymore, they're settling live payments on them. This is exactly the "incumbents absorb the plumbing and keep the customer" outcome we've been steel-manning.

3. Circle buys Tazapay for $400M and launches its own chain the day after the vote

Circle spent the week doing two things at once: buying distribution and shipping infrastructure.

On the deal: Circle agreed to acquire Tazapay, a cross-border B2B payments firm, for $400 million in an all-stock transaction. Why it matters in numbers: Tazapay connects payment companies and banks to more than 60 banking and fintech partners across 100-plus markets, processes $25 billion-plus in annualized payment volume, and roughly 60% of its transactions already involve stablecoins, a ready-made pipe to push USDC through (Thinking Crypto, Sep 9, pundit, reporting the filing).

On the infrastructure: Circle's ARC blockchain launches on mainnet September 16, the day after the Senate vote. Circle's chief economist and head of research Gordon Liao (operator/insider) gave the clearest explanation yet of why a company that already runs USDC on other people's chains would build its own. His answer, in plain terms: traditional institutions need guarantees today's blockchains can't provide.

"We built ARK around the principles for financial market infrastructure, which is a guideline that Basel had put out for payment companies and major financial institutions… A lot of these institutions require a certain level of liveness and certain level of finality that is beyond what current blockchains could address."

USDC is the chain's built-in fee token (so a user holding USDC "doesn't need to acquire any other token"), and the design targets settlement of "not just the current amount of crypto volume, but trillions of dollars of real world assets." Aave will be live on day one, offering USDC borrowing at roughly 3.9% with a 2% rebate (Paul Barron, Sep 11).

Liao also pushed back, gently, on the lazy version of the Circle bull case. Asked if Circle just makes more money when yields rise, he said: "I want to dispel the myth that stablecoin issuers primarily only make money from the interest… the circulation also matters." A useful reminder for anyone modeling CRCL purely off the fed-funds path.

And he tied the whole thing to the T-bill story we track. Quoting his boss Jeremy Allaire's forecast of a $3 trillion stablecoin market by 2030, Liao's co-host did the math: turning over short-dated collateral every couple of years would make stablecoins "a bigger buyer of sovereign debt than China, Japan, the UK, and Saudi Arabia combined." Liao's framing, that this is "very much aligned with what the U.S. Treasury is already doing" by issuing more short-term debt to sticky buyers, is the reserve-demand flywheel stated by an insider (DeFi Decoded, Sep 10) [DIRECTIONAL, 2030 forecast].

4. Coinbase arms the community banks, right before the vote

In the single shrewdest political move of the week, Coinbase tried to knock the legs out from under the bank lobby's best argument. The banks' scare story is that stablecoins will trigger deposit flight from small-town banks. So days before the vote, Coinbase partnered with MOV to give community banks and credit unions stablecoin capabilities, acceptance, settlement and real-time funding, first reported by CNBC. Coinbase already provides this kind of digital-asset infrastructure to JPMorgan, PNC and Citi; now it's extending the same hand to the community banks the lobby claims to be protecting (Thinking Crypto, Sep 11).

The politics got personal. Community banker Jill Castilla posted that "not talking will be the demise of clarity," implying the crypto side won't collaborate. White House crypto director Patrick Witt fired back publicly, calling the framing "misleading" and posting receipts, texts showing bank representatives had already been to the White House, and noting the stablecoin-rewards compromise struck by Senators Tillis and Alsobrooks "is the result of collaboration with the community banks." Translation: the industry says it already gave the banks a deal, and the banks came back for more (Thinking Crypto, Sep 11).

Why it matters: whether or not it saves the vote, Coinbase just repositioned itself as the picks-and-shovels supplier to both sides of the disintermediation debate, the challengers and the incumbents they're supposedly disrupting.

5. The quiet consolidation: neobanks are selling themselves, and the yield math keeps compounding

Away from the Beltway drama, two operator interviews told you where the money actually is, and where the bodies are.

Sam Kazemian (founder of Frax, operator) and Raagulan Pathy (founder of neobank Kast, operator) explained why the stablecoin neobank land-grab is turning into a shakeout. Pathy, fresh from briefing senior SEC staff, framed the demand side bluntly: "the world essentially wants to bank in America… There is unlimited demand for that. There's massive, massive demand." But the supply side is crowded, and Pathy said he now gets roughly two M&A inbounds a day from neobanks "throwing their hands up." Kazemian's survival thesis: brand isn't a real moat; scale and exclusive partner deals are. "My thesis is that it's a go big or go home. So I believe [Kast] would be a $100 billion company or we will fail. I don't think there's any in between" (The Rollup, Sep 13). The episode's title, "Why Neo Banks Are Quietly Selling Themselves", is the read-through: expect a wave of consolidation, and expect the winners to eat the deposit margin the incumbents are fighting to keep.

Meanwhile, the ZeroHash CEO (operator) revealed the firm won Morgan Stanley's crypto business and walked through the B2B use case that's already live: ZeroHash powers Gusto's global payroll, so a freelancer in the Philippines gets paid instantly instead of waiting three-to-five days on a SWIFT wire, and a small business can fund payroll "just-in-time" instead of pre-funding seven-to-ten days ahead. Given that payroll is "60% to 70% on average" of a small business's total costs, that's a real working-capital unlock, not a demo. His framing of the market: two billion "nodes" globally can now receive stablecoins, and the industry is fragmenting "on-chain money" into legally distinct buckets, GENIUS-defined stablecoins, tokenized deposits (JPMorgan's approach), tokenized money-market funds, and CBDCs (which "won't exist in the U.S., at least under this administration") (The Fintech Blueprint, Sep 10).

The debate

Do regulated stablecoins genuinely disintermediate banks, or do the banks and networks simply co-opt the tech and keep the value? This week both sides scored.

The co-option case had the better week, and it wasn't close on the infrastructure front. US Bank ran a live stablecoin payment. DBS and Citi settled tokenized deposits on SWIFT. Circle, the supposed disruptor, built its new chain on Basel's banking-infrastructure principles and stocked its validator set (from prior weeks) with Visa, Mastercard, BlackRock and DTCC. Coinbase is wiring community banks into stablecoin rails. Every one of these is an incumbent (or an incumbent's chosen vendor) absorbing the plumbing rather than being bypassed by it. The banks fought the technology, lost, and are now adopting it on their own terms, wholesale settlement first, where the deposit franchise is defended, not disrupted.

The disintermediation case, though, is quietly winning the economics, and the calendar. The whole point of Section 404 is to stop US payment stablecoins from paying interest, because a dollar that pays you is strictly better than a checking account that doesn't. That clause rides on a vote the vote-counters put at 5%. If it fails, yield-bearing and reward-heavy dollars keep scaling in the open, Ethena's USDe is already the fifth-largest stablecoin at ~$4.4 billion (Thinking Crypto, Sep 9), and the neobank founders describe genuinely unlimited global demand for a dollar account that a community bank can't offer. And ZeroHash's payroll story shows the disintermediation isn't hypothetical: it's already cheaper and faster than the correspondent-banking rail for cross-border money.

The tension in one line: the banks are winning the wholesale settlement layer decisively, while the yield-bearing challengers eat the retail deposit margin from underneath, and Washington may not close the loophole that lets them.

My read: this was a co-option week on the surface and a disintermediation week underneath. The banks going live is the louder headline, but it's mostly about defending institutional float and cutting their own settlement costs, not winning your checking account. The thing that actually decides whether stablecoins eat deposits is Section 404, and Section 404 is about to be a non-event on Tuesday. Watch the coexistence outcome harden: incumbents own the plumbing, challengers own the margin.

Stocks in play

Operator/insider commentary is flagged; everything else is investor or pundit color. Where a name was quiet, I say so.

  • CRCL (Circle), Loud week. Bought Tazapay for $400M (adds a $25B/yr, ~60%-stablecoin B2B pipe); ARC mainnet launches Sep 16 with Aave live day one. Chief economist Gordon Liao made the institutional-infrastructure pitch and reminded the market that circulation, not just rates, drives revenue. Bull: owning the rails plus buying distribution; the $3T-by-2030 reserve-demand story; ARC's validator moat (Visa/Mastercard/BlackRock/DTCC from prior weeks). Bear: if Section 404 dies with the vote, uncapped yield-bearing rivals undercut Circle's ~50-60% reward-payout model; a rate cut expected at the Sep 16-17 FOMC pressures reserve income. Watch: the Sep 16 ARC launch and early asset/validator traction (DeFi Decoded, Sep 10; Thinking Crypto, Sep 9).

  • COIN (Coinbase), Loud week. Partnered with MOV to arm community banks with stablecoin rails days before the vote; already serves JPM, PNC and Citi. Bull: positioned as the neutral infrastructure layer for both incumbents and challengers; strong policy muscle. Bear: revenue leverage still tied to a Circle payout split that Section 404's fate could reprice. Watch: the Sep 15 vote and any community-bank onboarding numbers (Thinking Crypto, Sep 11).

  • V (Visa), Covered. Now supports 160 stablecoins in its linked-card programs; stablecoin settlement volume topped a $20B annualized run-rate, up 15x year-over-year, and Visa wants blockchain lenders to underwrite issuers off VisaNet data. Bull: positioning to settle stablecoins, not be bypassed. Bear: on-network settlement still compresses interchange over time. Watch: stablecoin-linked card volumes (Paul Barron, Sep 8; Thinking Crypto, Sep 9).

  • MA (Mastercard), Quiet-ish. No standalone Multi-Token Network update this week; named only as a Circle ARC validator (prior weeks). Watch: any MTN metrics.

  • C (Citi), Covered, insider. Settled a live weekend tokenized-deposit payment with DBS on SWIFT's blockchain ledger; also a Coinbase infrastructure client. The clearest example this week of a money-center bank actually transacting on the new rails. Watch: volume on SWIFT's ledger and whether Citi Token Services scales (Thinking Crypto, Sep 8).

  • JPM (JPMorgan), Referenced, light. Named as the archetype for tokenized deposits (its Kinexys/JPMD approach) versus stablecoins, and as an existing Coinbase infrastructure client, but no fresh Kinexys or Dimon commentary. Watch: whether JPMorgan joins a bank consortium coin or stays solo (The Fintech Blueprint, Sep 10).

  • MS (Morgan Stanley), Covered, indirect. ZeroHash disclosed it won Morgan Stanley's crypto business, a data point after weeks of MS silence, though we heard it from the vendor, not the bank. Watch: what MS actually launches on that infrastructure (The Fintech Blueprint, Sep 10).

  • BK (BNY), Light. Named as custodian for Ripple's RLUSD reserves under its trust-charter model. Watch: custody mandates as more bank-grade stablecoins launch (On The Chain, Sep 12).

  • Tether (USDT), Covered. Pushed into private credit with a $400M fund alongside Fasanara, backing an evergreen "Stable Fund" vehicle aiming to raise up to $3B from outside institutions, Tether sourcing loans and providing stablecoin payment rails. The float is being put to work well beyond T-bills. Watch: foreign-issuer treatment under any Clarity/GENIUS follow-through (Thinking Crypto, Sep 11) [DIRECTIONAL, fundraising target].

  • HOOD (Robinhood), Loud week, but on tokenized stocks, not stablecoins. The memecoin-vs-tokenized-equity experiment kept stress-testing regulators: a "Boner" coin backed by Hims traded at 3-4x Hims's actual Friday close over a weekend; roughly $13.4B of tokenized stock tokens are outstanding and unregistered in the US; AMC CEO Adam Aron called them "vile" and "contemptible," and Robinhood's legal chief Dan Gallagher (a former SEC commissioner) told regulators "we will not desist." Hood chain is now the top gas payer on Ethereum with $5M+ daily fees. Bull: distribution is capturing crypto's value. Bear: this is inviting a securities-law fight. Watch: any SEC/CFTC comment on tokenized-equity integrity (Empire, Sep 11; Unchained, Sep 9).

  • GLXY (Galaxy Digital), Light. Galaxy's own podcast carried on-chain-finance coverage; no fresh company data point this week (Galaxy Brains, Sep 10).

  • MoneyGram, Covered, light. Launched a USDC-powered Visa card in Colombia, legacy remitter, new rails (Thinking Crypto, Sep 11).

  • Frax / Kast (private), Covered, insider. Founders Kazemian and Pathy flagged an accelerating neobank consolidation wave and "unlimited" global demand for dollar stablecoin accounts; Frax leaning on exclusive partner deals as its moat. Watch: announced M&A among stablecoin neobanks (The Rollup, Sep 13).

  • QUIET this week (no in-window podcast coverage): PYUSD/PayPal, XYZ/Block, FI (Fiserv), FIS, GPN (Global Payments), SOFI, GS (Goldman Sachs), BAC (Bank of America), WFC (Wells Fargo), Anchorage, Fireblocks, BitGo. Two silences stand out. First, the payment-processor middle tier, Fiserv, FIS, Global Payments, is now on a multi-week silent streak, and this week ZeroHash pointedly used "the last time you used a Fiserv or WorldPay interface" as the before picture. Second, PayPal's PYUSD went unmentioned during a week when US Bank, Citi, Circle, Tether and a neobank shakeout all made stablecoin news. Silence duly on the record.

Read-throughs

  • Card networks / interchange: Co-option, continued. Visa now supports 160 stablecoins in card programs and is settling $20B a year (15x growth), but it's doing so by becoming the settlement layer, not the disrupted one. The near-term read defends the franchise; the long-term read is that on-network stablecoin settlement and yield-bearing dollars still hollow the interchange pool. Toll booth or bypassed road, still unresolved, but Visa is aggressively building the toll booth.

  • Money-center & correspondent banks: This is the week the megabanks stopped talking and started transacting, US Bank live on Stellar, Citi live on SWIFT's ledger with DBS. The prize is cheaper wholesale settlement and defended institutional float, not retail deposits. Correspondent banking is the clearest casualty: ZeroHash's Filipino-payroll example (instant vs 3-5 days) is exactly the cross-border margin that flows through correspondent rails today.

  • Payment processors: The quietest and most exposed corner. Fiserv, FIS and Global Payments said nothing again while ZeroHash, Circle (via Tazapay), Coinbase (via MOV) and Visa wrote the stablecoin distribution playbook around them. The silence is starting to read as a competitive gap, not a lull.

  • Custody / exchange infrastructure: Picks-and-shovels keep winning regardless of which coin wins. ZeroHash bagged Morgan Stanley; Coinbase added community banks to a client list that already includes JPM, PNC and Citi; BNY sits as RLUSD custodian. The infrastructure providers monetize the co-option and the disintermediation.

  • Treasury-bill demand: Structurally reinforced by an insider this week. Circle's Gordon Liao explicitly tied stablecoin reserves to "sticky" front-end demand that lets Treasury issue more short-term debt, and endorsed the $3T-by-2030 math that would make stablecoins a larger buyer of sovereign debt than China, Japan, the UK and Saudi Arabia combined. With a rate cut expected at the Sep 16-17 FOMC, note the tension: lower yields squeeze issuer income even as reserve balances keep growing the price-insensitive T-bill bid.

What changed vs last week

  • The catalyst arrived, and it's a fizzle. Last week the Sept 15 vote was the event on the horizon at ~14% odds. This week it's tomorrow, the vote-counters have it at 5% (prediction markets ~16%), and the bank lobby (ABA, ICBA) formally escalated the yield fight on Sep 11. The disintermediation catalyst we've circled since July is about to whiff in public.

  • Section 404 got re-litigated at last. For weeks the podcasts argued the vote's odds but not the clause. This week Ron Hammond named it directly: the entire remaining fight is stablecoin yield, and the banks are trying to pick off crypto-naive senators to kill it. The clause is unchanged, but it's now openly the crux.

  • Co-option went from announcement to execution. Last week it was a 21-bank consortium with a 2027 date. This week it's live transactions, US Bank on Stellar, Citi/DBS on SWIFT, plus Circle buying Tazapay and launching ARC. The incumbents are transacting, not just planning.

  • Coinbase changed the political board. New this week: Coinbase arming community banks days before the vote, directly attacking the bank lobby's deposit-flight argument, with a public White House pile-on from Patrick Witt. Whether or not it saves the vote, it reframes Coinbase as neutral infrastructure to both camps.

  • A consolidation signal appeared. New this week: operators (Kazemian, Pathy) describing a neobank shakeout, "quietly selling themselves," ~2 M&A inbounds a day. The land-grab phase is ending; watch for deals.

Bottom line vs last week: the vote we've been waiting for is here and looks lost, the yield clause is finally the explicit battleground, and the banks answered the whole debate by simply going live. Two months ago the question was "will the yield clause survive the vote?" Tomorrow the question becomes "the vote failed, so who captures the margin while Washington isn't looking?"