Newsletter · · Ashutosh Agarwal

Twenty One Banks Build a Joint Stablecoin as Clarity Act Odds Crater - Stablecoins Eat Banking - Week of September 7, 2026

Stablecoins Eat Banking for the week of September 7, 2026: a consortium of 21 global banks confirmed a joint dollar stablecoin for 2027, the odds of the Clarity Act passing this year cratered to roughly 14 percent, and Ethena and Circle kept scaling the yield-bearing and infrastructure plays regardless.

Stablecoins Eat Banking

Week of September 7, 2026: Twenty One Banks Build a Joint Stablecoin as Clarity Act Odds Crater

Forty-seven relevant podcasts landed inside the seven-day window, and two of the stories they carried were each big enough to headline a normal week. First, twenty-one of the largest banks on earth confirmed they are building their own stablecoin. Second, the September 15 crypto-market vote everyone was counting on to pass now looks like it probably will not. One of those is the co-option thesis turning into a press release. The other is the disintermediation catalyst slipping through our fingers.

TL;DR

  • Wall Street is building its own stablecoin, for real this time. A consortium of 21 banks including Goldman Sachs, Bank of America, Citi, Deutsche Bank and UBS confirmed a joint US-dollar stablecoin targeting a first-half-2027 launch, with a euro version to follow. The group has doubled from 10 banks (October 2025) to 21. This is the incumbent co-option thesis made concrete, and even crypto commentators think the banks will still lose the retail market to Tether and Circle (Daily Crypto News, Sep 2; The Wolf Of All Streets, Sep 4).
  • The September 15 Clarity Act vote is now expected to fail. Odds of the bill being signed into law this year have cratered to roughly 14%, and a respected crypto trader put it at 10%. A new wrinkle: even if the Senate passes it, the House cancelled its late-September sessions, so it may not reach the President's desk in 2026. That is a sharp reversal from last week, when Coinbase's CEO was forecasting "over 60 votes" (Discover Crypto, Sep 1; Galaxy Brains, Sep 3; Thinking Crypto, Sep 4).
  • The yield workaround is scaling in the open. Ethena's CEO laid out economics that put its yield-bearing dollar (USDe) at a 6–8% collateral yield versus 3.5% for Tether and Circle, and launched a stablecoin neobank paying "the best cashback in the market." Meanwhile Circle's own layer-one network, with Visa, Mastercard, BlackRock, DTCC, MoneyGram and Standard Chartered as validators, goes live September 16 (On The Brink, Sep 2; This Week in Fintech, Sep 2).

What's new

1. Twenty-one banks, one stablecoin, and a 2027 launch date

This is the story of the week. A consortium of 21 major financial institutions, with named participants including Goldman Sachs, Bank of America, Citi, Deutsche Bank and UBS, has committed to forming a new joint company to launch a US-dollar stablecoin, targeting go-live in the first half of 2027. The project began in October 2025 with 10 banks and has more than doubled since. The group is also planning versions in other G7 currencies, with the euro named as the priority (host Matt on Daily Crypto News, Sep 2; host Tony Edward on Thinking Crypto, Sep 3).

Two details make this more than a headline. It will be built on public blockchains (which chain is not yet announced), not a walled-garden private ledger, a meaningful concession to how the technology actually works (host Scott Melker on The Wolf Of All Streets, Sep 4). And it sits alongside two other consortia that surfaced in recent weeks: OpenUSD, a fintech group led by Stripe and Coinbase building on Stripe's Tempo chain, and a coalition of 39 US regional and community banking associations planning their own network so they aren't squeezed between the megabanks and crypto (The Wolf Of All Streets, Sep 4).

Why it moves the thesis: the incumbents have stopped fighting the technology and started adopting it. As one host framed the capitulation, the banks "fought as much as they could" to block the neobanks and kill the rewards, "and now they've gotten to a place where they have no choice but to compete" (Onramp Bitcoin Media, Sep 3). The open question, steel-manned below, is whether a "Chase-or-Goldman coin" wins anyone over.

2. The Clarity Act vote is cratering, and it's the whole ballgame for pricing

Last week the September 15 vote was a coin flip. This week the market has all but written it off. The Senate returns to session on Tuesday, September 15 and holds a cloture vote (a procedural vote needing 60 of 100 senators to advance the bill). The odds of the Clarity Act actually being signed into law this year have dropped to about 14%, on prediction-market-style pricing rather than a hard count (hosts Kelly and Drew on Discover Crypto, Sep 1).

The most credible voice on this was Dan Matuszewski, co-founder of crypto trading firm CMS Holdings and a former head of trading at Circle, who put it bluntly on Galaxy's own podcast:

"I think the odds are pretty low here now, to be honest. I've publicly said it was 10% odds now because it didn't get done before the August recess... they come back into session on Tuesday, September 15th, and I think they leave the first week in October. So you're talking about two and a half weeks, basically."

He added the framing that matters most for a book: with odds this low, "clarity now is pretty well priced... price to perfection," which means "the downside impact of a failed vote is relatively capped, and you're looking at a possible huge upside catalyst with minimal downside if nothing happens" (Galaxy Brains, Sep 3). In other words, the asymmetry is now to the upside, because the market already assumes failure.

A genuinely new obstacle appeared this week. Because the Senate rewrote the bill (it started life in the House as FIT21, became the Clarity Act), it has to go back to the House for a final vote before reaching the President. But House Speaker Mike Johnson has cancelled the late-September House sessions, so even a Senate win might not reach Trump's desk in 2026. The consolation, per the same host: "if we just get it out of the Senate, that will be a huge win" (Tony Edward, Thinking Crypto, Sep 4).

Not all the news was bad. The National Sheriff's Association dropped its opposition and went neutral, the last holdout law-enforcement group, which Congressman Tom Emmer publicly thanked them for (Tony Edward, Thinking Crypto, Sep 6). And SEC Chair Paul Atkins, the closest thing to an insider voice here, went on Fox News to say the bill "will be voted on in the Senate on the 15th of September... I anticipate and hope that it will be passed by the Senate and sent ultimately to the president's desk for a signature" (Discover Crypto, Sep 2). Bitcoin reportedly popped about $100 the moment he said it.

One point on continuity: the podcasts this week debated the odds of the vote, not Section 404 specifically, the yield-ban clause that has been the fulcrum of this series. That thread is unchanged, but the vote it rides on just got materially less likely. If the vote fails or slips, Section 404 doesn't get decided on September 15 at all, and the "hold-to-earn" question stays open by default.

3. Ethena is scaling the exact yield workaround the banks fear, in the open

While Washington argues about whether a stablecoin can pay you, Ethena is already paying people. Founder and CEO Guy Young gave the cleanest description yet of the economics of a yield-bearing dollar. His product, USDe, generates yield through "delta-neutral hedging," holding crypto and shorting an equal amount of futures so the price exposure cancels out, leaving the funding yield. He laid the whole spectrum out plainly:

"Tether earning 3.5% but paying out zero. Circle earning 3.5% and paying out roughly 50 to 60% blended on average. And then you have something like Ethena, which through the cycle has achieved something that's closer to 6 to 8% as an average yield. And then your cost of liabilities is either 3.5% or slightly higher."

That is a bank's net interest margin, described in one breath. Ethena scaled from zero to $15 billion in supply in under two years, is targeting $15–25 billion again over the next 18–24 months, and names Bybit, Binance, Robinhood, Coinbase and BlackRock among its distribution partners. He also noted that roughly 90% of non-equity tokenized products are held by just two entities, Ethena and Sky. These are issuer-reported figures from the CEO himself, so treat them as directional (On The Brink, Sep 2).

On a separate show the same day, Young unveiled Ethena Pay, a stablecoin-native neobank ("$7B crypto company," per the host's framing) offering what he called "categorically the best cashback in the market with the largest budget." The tell for anyone modeling issuer economics: he admitted most cashback programs "are like unsustainable in the long run" and "loss-making in terms of the cash you're paying upfront," funded by interchange fees plus partner cost-sharing. The revenue lines, in his words, are "interchange fees on payments, FX fees... and then the NIM generation." He claims a stablecoin back-end is "90 to 95% more efficient" than a Revolut-style bank, and that self-custody is "reg light," in that it "rips out the need for thousands of people sitting there within a regulatory and compliance function" (When Shift Happens, Sep 2).

Why it matters: USDe is not a GENIUS-Act payment stablecoin. It's the yield-bearing category that lives in the gap Section 404 is meant to close, and it's compounding regardless of what the Senate does.

4. Circle's own network goes live September 16, with Visa, Mastercard and BlackRock as validators

The day after the Clarity vote, on September 16, Circle launches ARC (branded "Arc"), its own layer-one blockchain, on public mainnet. A Circle executive walked through the design on This Week in Fintech: USDC is the native "gas" (the fee you pay to transact) at roughly a tenth of a cent, so a user who already holds USDC "doesn't need to acquire any other token."

The eye-opener is the validator set, the institutions Circle chose to actually run the network:

"We announced the foundational validator cohort, which includes BlackRock, DTCC, Mastercard, MoneyGram, Visa, Standard Chartered, Galaxy."

Read that list again. The card networks, the world's largest asset manager, the plumbing of US securities settlement (DTCC), and a global remitter are now co-operators of Circle's rails. Circle also raised $222 million in an ARK token presale in May, has 10-plus local-currency stablecoins already deployed on the chain (Mexican peso, Brazilian real, Korean won, Japanese yen, Australian dollar), and runs a cross-currency product called stableFX. For scale, the exec pegged the total stablecoin market at about $300 billion against roughly $3 trillion for all of crypto, and noted Circle's euro coin EURC is the largest euro stablecoin at "around 500 million" (This Week in Fintech, Sep 2). The competing chains named: Plasma (Tether-backed, zero-fee USDT, access to Tether's roughly 600 million users) and Tempo (Stripe-incubated, payments-focused).

5. Revolut launches a euro stablecoin, and the card networks are already inside

Revolut, at 80 million customers, launched EURR, a euro stablecoin, but it didn't build it. The coin is issued by Bridge (owned by Stripe) under Europe's MiCA regime, rolling out first in Denmark, Portugal and Poland. Notably, MiCA is strict enough that EURR carries no yield and no rewards. The whole euro-stablecoin market is only about $810 million, a rounding error next to roughly $300 billion in dollar coins, and Circle's EURC is about 65% of it. Revolut is also eyeing the UK regulator's sterling sandbox and US issuance once its US bank charter clears (Turnkey COO Hannah Arnold and Thomas Cowan of Bullish on Tokenized, Aug 31).

That episode is worth flagging for who hosts it: Kai Sheffield, head of crypto at Visa, and Simon Taylor of Tempo, Stripe's chain, both insiders to the thing they are describing. Sheffield's own pitch, in the show's sponsor read, is that Visa's stablecoin platform lets a business "mint, move and manage stablecoins" while remaining self-custodian, and spend the balance "anywhere Visa is accepted" via stablecoin-linked cards. The read-through: Visa isn't being disintermediated in this telling, it's positioning to settle the stablecoins. Also from the episode: Standard Chartered ($850 billion AUM) became the first bank-authorized distributor of Hong Kong's regulated HK-dollar stablecoin, starting with tokenized money-market-fund settlement in Q4 2026; and a stablecoin neobank called Facet raised $68 million led by Japan's SBI at a $1 billion valuation, claiming (unverified, as private-company figures) $40 billion in annual volume across 125 countries and 12 straight profitable months (Tokenized, Aug 31).

The debate

This is the cleanest week we've had for steel-manning both sides, because both sides showed up with news.

The disintermediation case got a real champion in Guy Young. If a self-custodial dollar can pay a genuine 6–8% collateral yield versus 3.5% at the incumbents, run on a back-end "90–95% more efficient" than a neobank, and scale globally on day one because it's "reg light," then the stablecoin isn't competing with a deposit, it's strictly better than one, and it takes the net interest margin with it (When Shift Happens, Sep 2). Tether's Paolo Ardoino makes the sovereign-scale version of the same argument (see Stocks in play). The whole point of Section 404 is to make this illegal for a US payment coin, and Section 404 rides on a vote the market now expects to fail.

The co-option case got its definitive artifact. Twenty-one of the biggest banks on earth are building a coin; Circle's new chain is validated by Visa, Mastercard, BlackRock and DTCC; Revolut rents its issuance from Stripe's Bridge; Standard Chartered distributes a bank-issued stablecoin. In this telling, the incumbents simply absorb the plumbing and keep the customer, the float and the settlement fees. Visa's own posture, "spend it anywhere Visa is accepted," is co-option stated out loud.

The pull-quote of the week, from a crypto host who is rooting for disintermediation: the 21 banks came together "because they do not want to be dominated by Tether and Circle, even though they will be dominated by Tether and Circle."

And that's the sharpest tension. Even the co-option evidence comes wrapped in skepticism that co-option will work on the retail side. As the Daily Crypto News host put it: "Can they win? I don't think they can... nobody wants to use, I don't know, Chase or Deutsche or Goldman stablecoin," while conceding the banks could absolutely win the institutional rails and cut their own costs (Daily Crypto News, Sep 2).

The read: this week resolved less than it looks. The co-option side is winning the infrastructure battle decisively, since the incumbents are all in and Circle has literally made Visa and Mastercard its validators. But the disintermediation side is quietly winning the economics battle at the retail edge, because a yield-bearing self-custodial dollar is a genuinely better product and Washington may not close the loophole on September 15. Watch for the coexistence outcome: banks own the wholesale and settlement layer, and yield-bearing challengers like Ethena eat the consumer deposit margin. Both can be true, and this week made both more true.

Stocks in play

  • CRCL (Circle), loud week. ARC/Arc layer-one goes live on public mainnet September 16, with a validator cohort of Visa, Mastercard, BlackRock, DTCC, MoneyGram, Standard Chartered and Galaxy, an extraordinary institutional moat if it holds. USDC is the native gas token, which structurally pulls USDC demand onto the chain. Bull: owning the rails, not just the token; the $222M ARK presale; EURC already the largest euro coin (roughly $500M). Bear: Section 404's failure to pass would let uncapped yield-bearing rivals (Ethena) undercut Circle's 50–60% payout model; a trader on Galaxy's show argued Coinbase "should have just taken that entire business... the fact they didn't is just wild" (This Week in Fintech, Sep 2; Galaxy Brains, Sep 3). Watch: the Sep 16 mainnet launch and early validator and asset traction.
  • COIN (Coinbase), covered, mixed. Named as an Ethena distribution partner and an OpenUSD consortium member. But investor Dan Matuszewski was pointed: "Coinbase, I think they need a change in management... the stock has been so bad over so long and they're getting their lunch eaten in multiple areas... they're letting Robinhood just run laps around them" (Galaxy Brains, Sep 3). Bull: distribution reach for every new dollar product; policy muscle (Armstrong lobbying hard on Clarity). Bear: the Robinhood pair-trade narrative is running against it. Watch: the Sep 15 vote and any Circle revenue-share commentary.
  • Tether (USDT), loud week. CEO Paolo Ardoino said Tether is already "top 5 buyer at the 3-month [Treasury] auctions" if you combine Cayman hedge funds, and expects to become a top-10 then top-5 Treasury buyer overall, framing USDT as "the decentralized ownership of US debt" spread across 650 million holders who "will not wake up all together one single morning" and sell, unlike an antagonistic nation-state. The Treasury-buyer ranking is self-reported, so treat it as directional. He's betting on AI agents transacting in USDT (investments in robotics firms Neura and G-Bionics; a Tether Wallet Development Kit). Tellingly, he personally does not think the Clarity Act will pass (The Wolf Of All Streets, Aug 31). Watch: foreign-issuer treatment and the annual audit.
  • V (Visa), covered by an insider. Visa is a foundational validator on Circle's ARC and Kai Sheffield (head of crypto) co-hosts Tokenized, pitching Visa's stablecoin platform and "spend anywhere Visa is accepted" cards. Bull: positioning to settle stablecoins rather than be bypassed. Bear: on-network stablecoin settlement still compresses interchange over time. Watch: stablecoin-linked card volumes.
  • MA (Mastercard), covered lightly. Named in Circle's ARC validator cohort. No standalone Multi-Token Network update this week, but being a validator on a rival's chain is itself a data point on co-option. Watch: any Multi-Token Network metrics.
  • GS (Goldman Sachs), BAC (Bank of America) and C (Citi), covered as consortium members. All three named in the 21-bank USD-stablecoin group targeting H1 2027. Citi and Goldman were also singled out as building a stablecoin network "on public blockchains" (The Wolf Of All Streets, Sep 4). Watch: which chain they choose, and whether the JV actually incorporates.
  • JPM (JPMorgan), referenced lightly. Named within the DTCC Digital Assets Working Group, and its FX and repo use cases were cited as a target for Circle's stableFX. No fresh Kinexys or Dimon commentary. Watch: whether JPMorgan joins the 21-bank coin or stays solo on Kinexys.
  • GLXY (Galaxy Digital), covered. A validator on Circle's ARC; Galaxy's own podcast carried the Matuszewski interview; SharpLink is using $100 million of staked ETH as collateral in the Galaxy Fund (This Week in Fintech, Sep 2; Paul Barron, Sep 2).
  • HOOD (Robinhood), loud week. Robinhood Chain is the retail story of the moment: roughly $1.6 billion in decentralized-exchange volume and $70–80 million in tokenized real-world-asset issuance, with memecoins pairing to tokenized stocks. Named an Ethena distribution partner. Bull: distribution "captures crypto's value," running laps around Coinbase. Bear: the memecoin-versus-tokenized-stock mechanics (a "Boner" coin dislocating tokenized HIMS to an implied $100-plus over a weekend versus a roughly $20 Friday close) could invite regulators to treat these as unregistered equity derivatives (Empire, Sep 4; Daily Crypto News, Sep 2). Watch: any SEC or CFTC comment on tokenized-equity market integrity.
  • Stripe, Bridge and Tempo, covered by insiders. Bridge (Stripe-owned) issues Revolut's EURR; Tempo (Stripe-incubated) is the OpenUSD consortium chain and a named competitor to Circle's ARC. Stripe is quietly one of the most important stablecoin distributors in the world. Watch: OpenUSD traction and Tempo's launch partners.
  • Fireblocks, Anchorage and BitGo, covered. Fireblocks (a Tokenized sponsor) claims "$100 billion in monthly stablecoin volume" powering Visa, WorldPay, Bridge and Revolut. All three are named custody-integration partners for Lido's institutional staking; BitGo's Mike Belshe was interviewed and BitGo acquired NYDIG's derivatives desk (Tokenized, Aug 31; Paul Barron, Sep 2).
  • MoneyGram, covered lightly. Named in Circle's ARC validator cohort, a legacy remitter buying into the new rails.
  • No in-window coverage: SOFI, PYUSD/PayPal, XYZ/Block, FI (Fiserv), FIS, GPN (Global Payments), WFC (Wells Fargo), MS (Morgan Stanley), BK (BNY Mellon). The payment-processor middle tier, Fiserv, FIS and Global Payments, again drew no in-window podcast attention, which stands out against how much everyone else in payments is saying. PayPal's PYUSD also went unmentioned in a week when Revolut, Circle, Ethena and 21 banks all made stablecoin news.

Read-throughs

  • Card networks and interchange: The co-option is now explicit. Visa and Mastercard are validators on Circle's ARC, and Visa's own pitch is "settle in USDC, spend anywhere Visa is accepted." Near term this defends the franchise (they capture settlement); longer term, on-network stablecoin settlement and yield-bearing dollars still hollow out the interchange pool. Watch whether the networks are the toll booth or the bypassed road.
  • Money-center and correspondent banks: The 21-bank coin is the tell. The megabanks concluded they can't beat stablecoins, so they're joining. The prize they're actually chasing is cheaper wholesale settlement and defending institutional float, not winning retail. A separate coalition of 39 regional and community banking associations is forming its own network precisely because it fears being crushed between the megabanks and crypto. The deposit-migration risk this series has tracked is now being hedged by the banks themselves.
  • Payment processors: Stripe (via Bridge and Tempo) and Visa are writing the stablecoin distribution playbook, while Fiserv, FIS and Global Payments have yet to put a comparable strategy on the record. That gap is starting to look competitive rather than incidental.
  • Custody and exchange infrastructure: Compounding continues. Fireblocks at "$100B monthly stablecoin volume," BitGo buying NYDIG's derivatives desk, and Fireblocks, Anchorage and BitGo all wiring into institutional staking. Picks-and-shovels names keep winning regardless of which coin dominates.
  • Treasury-bill demand: Structurally reinforced and getting louder. Ardoino's claim that Tether is a top-5 buyer at 3-month auctions, plus the macro framing that stablecoins are becoming a deliberate tool to push demand to the front of the Treasury curve, keeps the price-insensitive T-bill bid intact (Rabobank strategist Michael Every on TFTC, Sep 5). As long as reserve yield can't be passed to holders of payment coins, that coupon stays with issuers, and if Section 404 fails to pass, the yield-bearing category (Ethena) grows the bid from the other direction too.

What changed vs last week

  • Operator interviews carried the week. Forty-seven relevant podcasts landed in the window, including operator interviews with the CEOs of Tether, Ethena, FV Bank and BitGo.
  • The Clarity vote flipped from likely to unlikely. Last week the odds were a coin flip and Coinbase's CEO was forecasting "over 60 votes." This week the odds are roughly 14% (and a respected trader says 10%), with a new structural obstacle: the House cancelled its late-September sessions, so even Senate passage may not reach the President in 2026. This is the single biggest change in the standing thesis, since the disintermediation catalyst that has been circled on the calendar is now expected to whiff.
  • The co-option thesis got concrete. Last week it was a framework. This week it's a 21-bank consortium with a 2027 launch date, a Circle chain validated by Visa, Mastercard, BlackRock and DTCC, a Revolut euro coin, and a Standard Chartered distribution deal. The incumbents are unambiguously in.
  • The yield workaround stepped into the open. Ethena put hard numbers on the yield-bearing model (6–8% versus 3.5%) and launched a neobank, exactly the "use to earn" category Section 404 targets, scaling while the clause meant to govern it hangs on a failing vote.
  • Section 404 itself was not re-litigated. The podcasts argued the vote's odds, not the clause. Treat the Section 404 debate as standing and unchanged from prior weeks, but note the vote it depends on is now more likely to be a non-event on September 15.

Bottom line: the calendar catalyst weakened, the co-option evidence hardened, and the disintermediation economics got a louder, better-quantified advocate. Two weeks ago the story was "will the yield clause survive the vote?" This week it's "the vote may not happen, and the yield product is scaling anyway."