Nu Holdings is likely to change the terms of the stablecoin yield inside Nu Global before the end of Q1 2027. The Brazilian group launched in the United States on September 10, 2026 paying 3.50% APY on balances held in Circle’s USDC and 2.20% on balances held in EURC, roughly 130 days before the GENIUS Act’s statutory effective date of January 18, 2027. Three signals observed in September point the same way: Nu went live on someone else’s bank charter, the Federal Reserve only opened its GENIUS Act rulemaking on September 29, and the Comptroller’s charter throughput is concentrated in a lane Nu is not standing in. The call is deliberately narrow and checkable: Nu’s published US rate disclosure for USDC balances is likely to read differently on March 31, 2027 than it reads today.
In short
- The prediction: Nu likely cuts, recasts or relocates the 3.50% USDC yield inside Nu Global before March 31, 2027, either by trimming the rate, reframing it as a time-boxed reward funded from Nu’s own margin, or moving it onto the insured deposit leg held at Lead Bank.
- The clock: the GENIUS Act takes effect on the earlier of January 18, 2027 or 120 days after final implementing rules, and because no final rule can now precede that date, the statutory backstop governs.
- Signal 1: Nu launched US banking on September 10, 2026 through Lead Bank of Kansas City, Missouri, with Nubank, N.A. still holding only a conditional charter approval from late January 2026.
- Signal 2: the Federal Reserve published its GENIUS Act proposed rule on September 29, 2026 with comments closing November 30, which places finalisation well beyond the effective date.
- Signal 3: the Comptroller’s September 18 approvals went to uninsured national trust banks with no deposits and no federal insurance, a lane that does not help a firm whose product is a yield-bearing consumer balance.
Why this matters now
Nu is not a marginal entrant. The group serves more than 140 million customers across Latin America and has spent three years demonstrating that a mobile-first bank can take primary-relationship share from incumbents in markets where card economics are rich and branch networks are expensive. Its arrival in the United States is therefore read by the market as a competitive event rather than a novelty launch, and its US chief executive Cristina Junqueira framed the ambition plainly: “We want to earn the place of being people’s primary banking relationship.”
What makes the launch analytically interesting is not the ambition but the plumbing. Nu did not wait for its own national bank charter. It went live on a partner bank’s balance sheet and bolted a stablecoin account onto the side, and that stablecoin account is the piece carrying the headline yield.
The structure works today. Whether it works after January 18, 2027 is a genuinely open question, and the answer shapes how every other foreign bank and large fintech reads the US entry playbook over the next eighteen months.
For merchants and platforms the stake is indirect but real. A consumer balance paying 3.50% in a dollar token, movable across more than 35 countries at low cost, is a payment instrument, and the rules that decide whether it can pay yield also decide how much of the consumer wallet ends up sitting in tokenised form rather than in cards and bank deposits.
There is also a timing asymmetry worth naming early. Nu chose to launch in the quarter before a statutory compliance date rather than the quarter after it, which is either a deliberate land-grab funded by a planned product change, or an assumption that the rules will settle in the industry’s favour. Either reading is defensible, and both lead to the same observable test at the end of Q1 2027.
Signal 1: Nu launched on someone else’s charter, with the yield attached to the token
On September 10, 2026 Nu switched on US consumer banking. Deposits are held by Lead Bank, a Kansas City, Missouri institution insured by the Federal Deposit Insurance Corporation. The headline offer is 3.50% APY on the standard account, rising to as much as 4.50% APY on savings goals capped at $10,000 for customers who opt into the Nu Mastercard, alongside a credit card carrying unlimited 1.50% cash back and no annual fee.
Alongside that sits Nu Global, a multicurrency account in which customer deposits are converted into Circle’s USDC or EURC. Nu pays 3.50% APY on the dollar-token leg and 2.20% on the euro-token leg, supports transfers across more than 35 countries, and offers bitcoin and ethereum trading in the same app. Circle and Nu have been partners since December 2023, when USDC became the exclusive dollar stablecoin on Nubank’s Brazilian platform.
The revealing detail is sequencing. Nu filed its national bank charter application with the Office of the Comptroller of the Currency on September 30, 2025 and received conditional approval 121 days later, in late January 2026. Eight months after that approval, the US business launched on a third party’s charter rather than on Nubank, N.A.
Chief executive David Vélez made the sequencing explicit: “We began to work on this project with Lead Bank before we had decided to apply for a banking license.” Read as revealed preference, that is a company that built a partner-bank stack as the primary route and treated the charter as a later upgrade, not a launch gate. Firms that expect a charter imminently do not usually build, staff and market a full consumer stack on rented rails first.
The conditions attached to the approval reinforce the point. Nu must fully capitalise the bank within twelve months, must open within eighteen months of approval, and must still secure deposit insurance approval from the FDIC and holding-company approval from the Federal Reserve. Planned US hubs were named as Miami, the San Francisco Bay Area, Northern Virginia and the North Carolina Research Triangle, which reads like a multi-year build rather than a Q4 flip.
Signal 2: the Federal Reserve only started its GENIUS Act rulemaking on September 29
The GENIUS Act was enacted on July 18, 2025. Section 20 sets the effective date as the earlier of eighteen months after enactment, which is January 18, 2027, or 120 days after the primary federal payment stablecoin regulators issue final implementing regulations. That formula matters, because the second branch can only pull the date forward, never push it back.
For the second branch to bind, final rules would have needed to exist by roughly September 20, 2026. They do not. The Treasury published its proposed rule on payment stablecoin issuance, offer and sale on August 18, 2026 with comments closing October 19, 2026. The Federal Reserve published its own proposed rule on September 29, 2026, with comments closing November 30, 2026, and a companion proposal on application procedures for Board-supervised insured depository institutions seeking subsidiary approval on the same day and the same deadline.
A comment window that closes on November 30 leaves 49 days before the statute bites. Agencies do not normally digest a comment file, draft a final rule, clear interagency review and publish inside seven weeks over a holiday period. The practical reading is that January 18, 2027 is now locked as the effective date, a conclusion consistent with our earlier work on why the GENIUS Act effective date likely lands on January 18, 2027.
The substantive risk to Nu sits in the Comptroller’s proposal, published in the Federal Register on March 2, 2026 with comments closing May 1, 2026. That proposal goes beyond the statute by creating a rebuttable presumption that an issuer has breached the yield prohibition when it pays an affiliate or a “related third party” that in turn pays yield to stablecoin holders.
The definition is what should concern a distributor. A “related third party” is proposed to mean, in part, a person “offering to pay interest or yield to stablecoin holders as a service”, and separately any person on whose behalf or under whose branding the issuer issues stablecoins. Nu Global offers yield to USDC holders as a service. The presumption is rebuttable in writing, and the agency reserves case-by-case judgement where an arrangement functions as yield in economic substance, but the burden sits on the industry rather than the regulator.
Signal 3: the Comptroller’s throughput is in the uninsured lane, not Nu’s
On September 18, 2026 the Comptroller issued three corporate decisions granting preliminary conditional national trust bank paths: a conversion for Bastion Platforms, and de novo charters for Catena Trust Bank and Agora National Trust Bank. All three are uninsured. None takes deposits. None carries FDIC insurance, and their permitted activities run to custody, stablecoin issuance and redemption, issuer services and fiduciary advisory work.
The capital terms are modest by bank standards: a minimum of $6 million in tier 1 capital, of which the greater of half the tier 1 or $3 million must sit in eligible liquid assets, plus 180 days of operating expenses in eligible liquid assets for the first three years. The conversion carries a six-month consummation deadline. This is a fast, light lane, and it is where the agency’s September output went.
The volume context is striking. The Comptroller received 40 de novo applications in the preceding eighteen months, against 48 across the fourteen years from 2011–2024, and 23 of the 40 involve digital-asset activity in the business plan. The queue is congested and the congestion is concentrated in exactly the category that competes for the same supervisory attention Nu needs.
Nubank, N.A. is in the slow lane by construction. It is an insured, deposit-taking de novo with a foreign parent, which means the OCC decision is necessary but not sufficient: the FDIC must grant deposit insurance and the Federal Reserve must clear the holding-company structure. Two additional approvers, each with its own comment and analysis cycle, is a materially different timetable from an uninsured trust charter.
That asymmetry is not unique to Nu. The same structural gap runs through the US charter attempts by consumer credit fintechs, a pattern we traced in the BNPL bank charter race, where the applicants seeking insured deposit powers have consistently taken longer than those seeking narrower permissions.
The signals matrix
| Signal | Date observed | Source type | What it implies for the yield | Strength |
|---|---|---|---|---|
| Nu launches US banking on Lead Bank, with 3.50% APY on USDC in Nu Global | September 10, 2026 | Company launch and press coverage | The yield is attached to a token, not to an insured deposit Nu controls | High |
| Federal Reserve GENIUS Act proposed rule, comments close November 30 | September 29, 2026 | Federal Register proposed rule | Final rules cannot precede January 18, 2027, so the statutory date governs | High |
| OCC grants three uninsured national trust charters | September 18, 2026 | OCC corporate decisions | Agency throughput favours the lane Nu is not in; the charter escape hatch stays shut near term | Medium |
| OCC proposes a rebuttable presumption against related third-party yield | March 2, 2026 (comments closed May 1) | Federal Register proposed rule | Defines the specific legal exposure of a distributor paying yield on a token | Medium to high |
| Nu Mexico takes about fifteen months from licence approval to switching the bank on | April 2025 to August 6, 2026 | Regulatory milestone history | Establishes the group’s own base rate for licence-to-operations lag | Medium |
What the pattern suggests
Put the three September signals together and a single conclusion falls out. Nu has built a US consumer proposition whose most competitive feature, a 3.50% yield on a dollar token that moves across 35 countries, depends on a regulatory reading that the industry likes and the Comptroller has already proposed to narrow. The insurance policy against that reading going wrong would be Nu’s own charter, and the charter is not arriving in time.
The group’s own history sets the base rate. Nu Mexico received banking licence approval from the CNBV in April 2025, obtained its operations authorisation on July 9, 2026, and began trading as a fully licensed bank on August 6, 2026 after a technical migration the day before. That is roughly fifteen months from approval to switch-on with a single regulator, having operated in the interim as a Popular Financial Society with more than 10 million customers and $4.5 billion of deposits.
Apply that interval to the US and Nubank, N.A. lands in the second quarter of 2027 at the earliest, against an eighteen-month consummation window that expires around late July 2027. A 2027 opening is plausible. An opening before January 18, 2027 is not, which means the yield question has to be answered on the current structure.
The most likely resolution is undramatic. Nu probably does not withdraw from the US, and probably does not shut Nu Global. It most likely adjusts the disclosure: a lower headline rate, or a reward funded transparently from Nu’s own margin and described as promotional rather than as interest on the token, or a split in which the insured deposit leg carries the yield and the token leg carries the transfer utility.
That third option is the tidiest and the one worth watching hardest, because it reproduces in the US the same logic that drove the deposit-funding shift we documented in US buy-now-pay-later converting into deposit-funded banking. The cheapest compliant place to pay a consumer for their balance is an insured deposit, and the whole point of a charter is to own that place outright.
Prior precedents
| Case | Interim structure used | Approval milestone | Operations start | Lag |
|---|---|---|---|---|
| Nu Mexico | Popular Financial Society (SOFIPO) with 10m+ customers and $4.5bn deposits | CNBV licence approval, April 2025 | August 6, 2026 | About 15 months, single regulator |
| Nu United States | Partner bank (Lead Bank) plus a stablecoin account | OCC conditional approval, late January 2026 | Not yet; 18-month window closes around late July 2027 | Pending, three regulators |
| Uninsured trust charters (Bastion, Catena, Agora) | None required; no deposits taken | Preliminary conditional approval, September 18, 2026 | Six-month consummation deadline on the conversion | Months, single regulator |
The table makes the structural point better than prose can. The gap between the fast lane and Nu’s lane is not a few weeks of paperwork. It is the difference between one approver and three, and between a charter that cannot take a consumer deposit and one that can.
Wider context: the Coinbase-shaped hole and who actually pays
The statutory yield prohibition binds permitted payment stablecoin issuers. Circle is the issuer of USDC and EURC. Nu is a distributor, and on a plain reading of the statute a distributor paying rewards out of its own revenue is not the regulated party. This is the gap the industry has argued for since the bill passed, and it is why large exchanges have continued to pay rewards on stablecoin balances without obvious legal jeopardy.
The Comptroller’s proposal attacks the gap from the issuer side rather than the distributor side, which is a clever piece of drafting. If the issuer pays a related third party and that party pays holders, the presumption of breach attaches to the issuer. The commercial consequence is that the issuer, not the distributor, carries the enforcement risk, and issuers respond by rewriting distribution contracts rather than by waiting to be examined.
That is the transmission channel to watch. Nu’s exposure is less likely to arrive as a supervisory letter addressed to Nu and more likely to arrive as a renegotiated distribution economics term sheet from Circle. Stablecoin distribution is funded out of reserve income, and if reserve income can no longer legally flow to a partner who pays holders, the partner’s rate has to fall or be refinanced from elsewhere.
The same squeeze is already visible in the consumer card layer, where the economics of tokenised balances have proved harder to scale internationally than the launch announcements implied, a pattern we set out in why stablecoin cards likely miss 100 countries by December. Yield and distribution breadth are funded from the same pool, and regulation that constrains one tends to constrain the other.
None of this makes Nu Global a bad product. A near-free multicurrency account with transfers across 35 countries is a genuinely strong proposition for the US-to-Latin America remittance corridor, which is the obvious strategic target. The argument here is only that the yield line item is the fragile part of the offer, not the transfer utility.
Scenarios and rough odds
| Scenario | What it looks like | Rough odds | How to check by March 31, 2027 |
|---|---|---|---|
| Reprice and relabel (base case) | The USDC rate falls, or is recast as a promotional reward funded from Nu’s margin with an end date attached | ~55% | Nu Global’s published US rate disclosure and terms page |
| Migrate the yield to the insured leg | Yield is paid only on the Lead Bank deposit balance; the token balance becomes a transfer and holding instrument | ~20% | A split in the rate table between deposit APY and Nu Global |
| Hold the line | Nu keeps 3.50% and relies on being an unaffiliated distributor outside the statutory prohibition | ~20% | No change to the published rate after January 18, 2027 |
| Charter arrives first | Nubank, N.A. consummates, deposits move in-house and interest is paid as ordinary deposit interest | ~5% | An OCC licensing action recording consummation |
The base case and the migration case both count as the prediction landing. The hold-the-line case is the clean falsification, and it is far from negligible at roughly one chance in five.
Implications for merchants, issuers and platforms
For merchants the near-term effect is small and the medium-term effect is a question about where consumer float sits. If yield-bearing tokenised balances survive in consumer apps, a slice of the wallet migrates out of cards and into instruments with different acceptance economics. If they do not, the tokenised consumer balance stays a cross-border transfer tool and the card rails keep the float.
For issuers the renegotiation risk is the story. Distribution deals signed on the assumption that reserve income can be shared with partners who pay holders may need rewriting before January 18, 2027, and partners with thin margins will feel it first. Nu is not thin, which is one reason it may choose to absorb the cost for a period rather than cut the rate outright.
For platforms and payment service providers the practical read is to expect the menu of consumer-facing stablecoin propositions to shorten rather than lengthen through 2027. That narrowing mirrors what we expect on the acceptance side, set out in why US merchant stablecoin menus likely narrow before mid-2027, and the two dynamics are driven by the same rulemaking.
For competitors watching Nu’s entry as a template, the lesson is more nuanced than “launch on a partner bank”. The partner-bank route works for speed. It does not give you the one thing that makes paying customers for their balances unambiguously lawful, which is a deposit-taking charter of your own.
The corollary for anyone planning a 2027 US entry is to run the regulatory clock backwards from the compliance date rather than forwards from the launch date. Nu appears to have launched into a 130-day window, which is workable if the plan always assumed a product change at the end of it, and awkward if it did not.
One more implication deserves stating for anyone modelling the group. If the yield is repriced, the customer-acquisition cost of the US business rises rather than falls, because the 3.50% headline is doing work that marketing spend would otherwise have to do. Nu has the balance sheet to absorb that, having invested more than .4 billion into Mexico before that market turned, so the likelier outcome is a slower, cheaper US ramp rather than a retreat.
Caveats: what could go wrong with this call
The strongest counter-argument is the simplest. The statutory prohibition binds issuers, Nu is not an issuer, and the Comptroller’s presumption is a proposal that has not been finalised and that industry comments have pushed hard against. A final rule that expressly carves out unaffiliated third parties paying rewards from their own funds would leave Nu Global’s 3.50% intact and this prediction wrong.
Second, enforcement rarely arrives on the effective date. With no final rules in place on January 18, 2027, supervisors face a practical choice between forbearance and improvised action, and forbearance is the historical default. Nu could reasonably keep the rate through a quiet first quarter and change it only when a final rule lands.
Third, the resolution test has an ambiguity problem. If Nu trims the rate from 3.50% to 3.25% in response to falling short-term dollar rates rather than to the statute, an observer could score the prediction as landing for the wrong reason. The honest test is whether the structure of the disclosure changes, not merely the number.
Fourth, Nu could surprise on the charter. The Comptroller processed the application in 121 days, which is fast by historical standards, and a commissioner keen to show that the charter pipeline is open might push the remaining approvals faster than the Mexican precedent implies. That would not arrive before January, but it could arrive before the second half of 2027 and change the framing entirely.
Fifth, Congress remains a live variable. Amendments clarifying the scope of the yield prohibition have been floated by both the bank lobby, which wants it read broadly, and the digital-asset lobby, which wants it read narrowly. A legislative clarification in either direction would settle the question faster than any rulemaking.
Finally, this analysis rests on publicly reported product terms. Nu’s disclosures may already contain promotional language or rate-change rights that make a repricing routine rather than newsworthy, in which case the prediction lands but tells you less than it appears to.
Frequently asked questions
What exactly is being predicted?
That the terms on which Nu Global pays yield on USDC balances are likely to change before March 31, 2027, through a rate cut, a reframing as a promotional reward, or a shift of the yield onto the insured deposit leg held at Lead Bank. The transfer functionality and the multicurrency account itself are not expected to disappear.
Does the GENIUS Act actually ban Nu from paying yield?
Not directly. The statutory prohibition applies to permitted payment stablecoin issuers, and Circle rather than Nu is the issuer of USDC and EURC. The pressure comes indirectly, through the Comptroller’s proposed presumption that an issuer paying a related third party that pays holders has breached the prohibition, which pushes the risk into distribution contracts.
Why can the effective date not move?
Section 20 sets it as the earlier of eighteen months after the July 18, 2025 enactment or 120 days after final implementing rules. Since the Federal Reserve’s comment period alone runs to November 30, 2026, no final rule can be 120 days old before January 18, 2027, so the earlier of the two dates is the statutory one.
Is this not just a Coinbase problem that Coinbase has already survived?
That is the best counter-argument. Exchanges have paid rewards on stablecoin balances for years on the view that unaffiliated third parties fall outside the ban, and no enforcement has tested it. The difference in 2027 is that a rule with an explicit anti-evasion presumption, rather than statutory silence, will be in the field.
Could Nu simply get its charter and pay interest lawfully?
Eventually, yes, and that is plainly the plan. The obstacle is timing: conditional approval came in late January 2026, the FDIC and the Federal Reserve have not yet signed off, and the group’s own Mexican precedent took roughly fifteen months from licence approval to operations with only one regulator involved.
What would make this prediction clearly wrong?
Nu Global still advertising an open-ended 3.50% APY on USDC balances on March 31, 2027, with no promotional framing, no end date and no migration of the yield to the deposit leg. A final rule carving out unaffiliated distributors would be the most likely cause.
Does any of this matter to retailers and marketplaces?
Modestly, and mostly through where consumer float sits. A yield-bearing dollar token in a mainstream consumer app pulls balances out of cards and into an instrument with different acceptance and settlement economics, and the rules governing that yield therefore shape how quickly tokenised consumer money reaches checkout.
Is Nu’s US entry likely to succeed regardless?
The remittance and multicurrency proposition looks strong on its own merits, particularly across the US-to-Latin America corridor where Nu already has scale on the receiving side. The yield line is a customer acquisition lever rather than the core of the offer, which is one reason a repricing is more likely than a retreat.
When will we know?
Three checkpoints matter: the Treasury comment file closing October 19, 2026, the Federal Reserve comment file closing November 30, 2026, and the effective date on January 18, 2027. Nu’s published US rate disclosure on March 31, 2027 is the resolution test for this piece.