What sponsor banking does, and why different users choose it
connects a regulated bank with another company’s distribution, technology or customer interface. The end user may experience a single app, card or checkout flow even though the bank, fintech, processor and program manager have different responsibilities. Banking as a service (BaaS) is a broad commercial label rather than one uniform product or risk category.
For a household, the attraction can be a convenient transaction account, faster access to available funds or a payment-linked credit product. For a business, it can be account administration and money movement integrated with its working software. For a merchant or lender, sponsorship can provide the issuing, acquiring, disbursement or repayment connection behind a customer-facing service. These are possible product benefits; an announced capability does not show that customers adopted it or became financially better off.
The bank can gain fee income, deposits and access to customers it would find expensive to reach itself. The technology company can offer bank-provided services without first building or acquiring a bank. Both sides give up some economics and control. The commercial question is whether distribution and specialization create enough value to pay for servicing, losses, , technology and oversight throughout the relationship, including its exit.
Four models that should not be merged into one market statistic
A single partnership may include several of these activities. The relevant legal entity and product agreement determine the bank’s role; a fintech logo does not identify which party holds customer money or bears a loan loss. Direct bank–fintech integration and arrangements using middleware also have different record-access and continuity dependencies.
Scroll horizontally to see all columns.
| Model | What the bank supplies | Economic and risk boundary |
|---|---|---|
| Deposit/account program | Accounts and deposit holding, sometimes with debit cards or sweep arrangements | Funding value and service fees, offset by interest, support, compliance, reconciliation and correlated withdrawals |
| Card issuing/BIN sponsorship | Issuing relationship and network access; credit where the product includes lending | Net interchange and fees, with network/processor costs, rewards sharing, fraud and disputes; credit losses only where applicable |
| Payment/acquiring sponsorship | Access and settlement support for relevant payment flows | Fees versus returns, , sanctions controls, prefunding and settlement ; payment processing does not imply loan funding |
| Loan origination/credit sponsorship | Origination, temporary funding or retained lending under the program terms | Origination/servicing fees and retained spread, with credit, failed sales, repurchases and counterparty obligations that can outlast sale |
Who is using the model: verified relationships and their limits
The examples are a selected set of disclosed relationships, not a census of , customers or market share. A legal disclosure verifies a stated provider; a dated announcement verifies announced scope. Neither reveals confidential contract margins or establishes how every customer account is configured.
Scroll horizontally to see all columns.
| Relationship | Verified scope/status | What the evidence does not establish |
|---|---|---|
| Chime / The Bancorp Bank, N.A. and Stride Bank, N.A. | Chime names both banking providers and announced a Stride-parent acquisition agreement on September 8, 2026. [8][9] | The proposed purchase is not a completed change of ownership or proof that all customers use one bank |
| Mercury / Column N.A. and Choice Financial Group | Mercury’s legal page names both providers; the Column page links demand-deposit and debit agreements. Checked October 4. [10] | No customer count, program margin or independent service-quality outcome is supplied by this disclosure |
| Pathward / TabaPay | August 5 extension into 2031 covers acquiring sponsorship and money movement; some use cases serve Pathward lending partners. [11] | The announcement does not make every TabaPay payment a Pathward loan |
| Coastal / CCBX partners | CCBX is Coastal’s BaaS segment, with separate active and prelaunch/exit categories. [6] | Relationship counts are not unique users or proof all programs are profitable |
| FinWise / strategic lending and card programs | Lending originations and program fees are reported separately; July’s Tallied asset purchase brought card technology in-house. [7][14] | Technology ownership alone does not eliminate underwriting, servicing or integration risk |
Adoption is measurable, but the denominators are different
Chime reported 10.4 million Active Members at June 30, 2026, up 20% year over year, and $38 billion of Q2 Purchase Volume, up 17%. Its active definition requires member-initiated money movement in the period’s final calendar month; it is not a count of primary checking relationships. Purchase Volume means Chime-branded card purchases net of adjustments/refunds and excludes deposits, transfers and loan advances. These measures establish scale and usage for Chime, not industry penetration. [12]
Coastal reported 22 CCBX relationships in its Active category within 30 total at June 30: one testing, one onboarding, three letters of intent and three winding down. The wind-down category comprises relationships still active but preparing to exit. Its $4.56 billion quarterly loan sales included $3.68 billion of additional receivables from existing revolving-card arrangements. Those flows are not unique borrowers or a $4.56 billion ending portfolio. [6]
FinWise reported $1.6 billion of Q2 originations, versus $1.7 billion in Q1 and $1.5 billion a year earlier. Volume is a flow across its reported lending activity; it is not the same as the amount retained or the number of people served. [14]
An account count, active-member count, sponsor relationship count, payment dollar and loan dollar describe different stages of adoption. They cannot be added into a market total: the same person or transaction can appear in both a fintech disclosure and its sponsor’s results. Public disclosures reviewed here do not support a market-wide adoption rate, active-user retention curve or causal claim of improved financial inclusion.
Why payment scale and bank revenue are different measures
The Bancorp reported Q2 2026 card gross dollar volume of $53.453 billion and total fintech fees of $40.894 million. The volume definition covers spending on its issued prepaid, debit and credit cards. Total fintech fees span its fintech activities, so dividing those fees by card volume would not establish a clean card take rate. The group reported $8.41 billion average deposits at a 1.63% average cost; these are consolidated deposit measures, not a quoted funding price for each fintech partner. [13]
Analytically, the bank’s deposit benefit depends on average usable balances and the return available after needs, interest and servicing costs. A large quarter-end balance can be misleading if payroll or settlement timing creates a temporary spike. Swept deposits at other banks may create fees but do not provide identical on-balance-sheet funding value. A card program can generate substantial spending while the sponsor retains only a small contractual share of the revenue.
The fintech has a different income statement. Chime reported Q2 revenue of $670 million, including $430 million payments revenue, and $28 million GAAP net income. Its $260 revenue per Active Member is annualized quarterly revenue divided by average quarter-end Active Members, not quarterly profit per customer. These are fintech results, not the partner banks’ profitability. [12]
Gross loan yield is not the sponsor’s retained return
Coastal’s Q2 CCBX loan yield was 14.56%; after BaaS loan expense, its non-GAAP net loan-income yield was 7.07%. Both yields are annualized and based on average CCBX loans. The latter remains before funding and other costs. It reported a $42.1 million consolidated net loss and $68.8 million expense linked to one partner: $22.8 million provision plus a $46 million credit-enhancement-asset adjustment. Management described the issue as isolated; the figures do not establish a sector loss rate. [6]
The accounting boundary matters. Customer interest can enter bank revenue while payments to the partner appear elsewhere in expenses. A headline yield without that sharing cost overstates the economics retained by the bank. Conversely, fee income after a loan sale can continue even though the receivable no longer appears on the originating bank’s balance sheet.
FinWise’s Q2 Strategic Program fees were $5.31 million. It separately recorded $16.678 million of credit-enhancement income matched to provision on enhanced loans. That income reflects expected future recovery under contractual credit-loss support rather than ordinary customer fee revenue; recognizing it does not establish cash receipt. Availability of collateral and ultimate collection remain important. Its reported consolidated net income was $2.1 million. [14]
A contractual indemnity is a counterparty exposure. Its value depends on enforceability, claim priority, reserve control, replenishment and the partner’s ability to pay at the time a loss occurs. Correlation matters: the same deteriorating loan book can create a bank loss and weaken the party promising reimbursement. An unsecured recovery estimate and immediately usable cash serve different purposes.
What the published performance evidence can and cannot answer
The named results demonstrate customer activity and financial outcomes for particular companies and periods. They do not provide a controlled comparison between and directly delivered banking. Nor do the releases reviewed here give consistent program-level measures of customer retention, all-in customer cost, successful dispute resolution, fraud by transaction type or credit losses by common and borrower risk.
For lending, an origination-based loss percentage, annualized divided by average loans and a lifetime expected-loss provision are different measures. Credit-enhanced portfolios also require a distinction between gross borrower losses, bank-recognized provision, reimbursements received and unsecured enhancement assets. A low net loss after support does not establish low underlying defaults.
For consumer outcomes, fewer fees, convenience or faster payments are product characteristics until supported by observed results for a defined population and period. Marketing surveys and management-selected repeat-borrower comparisons can be informative, but they do not by themselves isolate product effects from customer selection. The evidence gap is especially large for customers who leave, are rejected or cannot access the service.
When partnering turns into ownership
Chime’s September 8 Form 8-K describes an agreement to acquire Central Service Corporation, Stride’s parent, for $590 million subject to adjustments. Chime targets the first half of 2027, subject to approvals and other conditions, and expects more than $100 million in net synergies. Those are announced terms and forecasts, not completed ownership or realized savings. [8][9]
FinWise’s July 20 Tallied announcement concerns an acquired technology platform and related assets already supporting its bank’s cards. It describes approximately $50 million of balances moving from credit-enhanced treatment to standard bank-retained balances, with related interest/interchange and credit exposure, and about $4 million of anticipated integration/transition costs over the following year, excluding acquired-intangible amortization. [7]
Analysis: integration can remove handoffs and revenue sharing while adding staffing, capital allocation, technology maintenance and direct risk responsibilities. The commercial comparison is therefore the combined return after those costs, not the gross amount of sponsor fees avoided. An acquisition changes who performs a function; it does not make the function unnecessary.
Worked economics: activity, contribution and stress
Illustrative monthly program, not a company result: 100,000 funded accounts generate $300,000 revenue, or $3 each. Processing costs $80,000, support $90,000 and fraud/disputes $60,000. Contribution is $70,000 before acquisition costs, fixed technology, compliance, profit sharing and tax. If revenue falls 20% and those costs do not change, contribution falls to $10,000. Accounts opened, funded accounts and economically active accounts are different populations.
A separate annual illustration produces the table below. It includes an internal capital/ charge for analysis; that charge is neither a GAAP expense classification nor a stated regulatory requirement. Revenue growth can coexist with falling contribution when complexity and losses rise faster.
Scroll horizontally to see all columns.
| Annual amount, $ millions | Base | Growth with strain |
|---|---|---|
| Revenue | 6.00 | 7.20 |
| Processing and servicing | (2.00) | (2.40) |
| Oversight and technology | (1.00) | (1.30) |
| Expected losses/remediation | (1.20) | (1.80) |
| Internal capital/liquidity charge | (0.60) | (0.72) |
| Contribution before tax/unallocated overhead | 1.20 | 0.98 |
A profitable program can still run short of cash
Continuing the annual illustration, a separate incident costing $3 million beyond the base loss allowance would exceed the $1.2 million contribution. If only $0.8 million of controlled cash collateral is immediately available to the bank, net incremental exposure is $2.2 million, turning the contribution into a $1 million loss. Later recoveries may improve the result without meeting today’s obligations.
Another hypothetical bank has $500 million of deposits, including $200 million sourced through one fintech: a 40% channel concentration despite many end customers. A 30% channel outflow costs $60 million. Available cash of $20 million plus $50 million of unencumbered securities monetized at a 10% haircut supplies $65 million, leaving $5 million before other obligations. Pledged securities or delayed access reduce that cushion. This is a illustration, not a regulatory-ratio calculation.
Loan sales create another timing gap. Origination can require bank funding before a purchaser pays, while eligibility disputes, market shutdowns or repurchase obligations can return risk to the bank. On-balance-sheet credit, temporary funding, servicing advances and contractual recourse are distinct exposures; sold principal alone does not describe the residual risk.
Customer funds: insurance, accurate records and access are separate
FDIC insurance applies to insured-bank failure, not a nonbank’s insolvency or an app outage. Money sent to a nonbank is not eligible until deposited at an insured bank and the relevant conditions are met. Pass-through coverage depends in part on records identifying ownership and each owner’s amount. A fintech’s bank relationship therefore does not insure every operational or counterparty failure. [5]
The Synapse disruption illustrates why this distinction matters. The FDIC’s September 2024 recordkeeping proposal expressly responded to uncertainty over actual deposit ownership, placement and access when a nonbank fails. This is cited as the agency’s historical explanation of the problem, not as a claim about current Synapse recoveries or a final rule. No current court-verified total of unrecovered customer funds is established in this review. [17]
An omnibus account can balance in aggregate while individual people’s balances are wrong. In a hypothetical $100 million pool, understating customer A by $50,000 and overstating customer B by $50,000 leaves the total unchanged. Reconciliation at customer and transaction level addresses a different question from agreement of the two aggregate totals. Gross unexplained differences and affected-customer counts reveal issues that a zero net break can hide.
The operational chain determines where problems reach customers
The agencies’ July 2024 statement describes risks from fragmented deposit operations and inadequate bank access to records; outsourcing does not reduce a bank’s compliance responsibilities. It reiterates existing principles rather than creating a new licence. [1]
An analytical reconstruction connects opening balance, authorized activity, settlement, pending transactions, fees and corrections to the closing customer position. Duplicate postings, reversals and migrations can create errors even when file totals agree. A direct right to receive data has limited practical value if the data are delayed, incomplete or unusable after a provider shuts down.
Analysis: using a second sponsor bank does not necessarily diversify a program if both banks depend on the same ledger, servicer or funding provider. Continuity also has a cost after commercial revenue stops: customer support, residual disputes, collections and account transfer or closure can require staff and funding throughout a servicing tail.
Scroll horizontally to see all columns.
| Operating dependency | Useful observable evidence | Customer/business consequence if it fails |
|---|---|---|
| Funds and beneficial ownership | Bank/settlement records linked to complete customer subledgers and explained exceptions | Delayed access, incorrect balances or uncertain insurance determinations |
| Fraud and identity controls | Losses by payment/product type, alerts, investigations, overrides and recovery timing | Unauthorized activity, account restrictions or unrecovered loss |
| Complaints and disputes | Case timestamps, assigned entity, applicable deadline and resolution history | Customers passed between firms or remedied too late |
| Product/credit changes | Approved terms linked to production configuration and deployment history | Unauthorized risk changes or inconsistent customer treatment |
| Provider continuity | Usable exports and demonstrated recovery/servicing transfer | An intact bank balance with no working customer service path |
Enforcement evidence needs both an original date and a current-status check
The Federal Reserve’s June 14, 2024 Evolve announcement identified AML, risk-management and consumer-compliance deficiencies following 2023 examinations. It expressly said the action was independent of Synapse’s bankruptcy. That limits the inference: the order is evidence of specified supervisory findings, not an adjudication of every dispute over missing funds. [15]
Evolve’s order, effective June 11, 2024, addressed fintech oversight, ledger responsibilities, complaints and controls. Paragraph 19 required prior supervisory approval for specified new Open Banking Division activities and relationships; paragraph 28 leaves provisions effective until changed in writing. No official termination was located in this review, but the reviewed record does not establish present compliance with each remediation requirement. A missing public termination is not proof that no remediation has occurred. [16]
Blue Ridge provides the contrasting status example: the OCC formally terminated its January 24, 2024 on November 13, 2025. The original action addressed BSA/AML and unsafe or unsound practices including capital, and IT controls. Describing that order as still active would be incorrect. Termination is a supervisory status change, not an industry-wide performance verdict or an assertion that every historical customer dispute has been resolved. [18][19]
Regulatory position checked October 4, 2026
The 2023 interagency third-party guidance addresses planning, diligence, contracting, monitoring and termination, with risk management proportionate to the relationship. It says it does not impose new requirements. This life-cycle framework is relevant to sponsor arrangements without making every vendor relationship equivalent. [2]
The September 11, 2026 agency announcement and September 15 Federal Register notice propose replacement third-party guidance, with comments due November 16, 2026. The agencies say they plan to rescind and replace existing guidance when finalized. As of this review, the cited documents are a proposal and comment process; this article does not treat them as a final replacement or new binding rule. [3][4]
Bank-specific enforcement orders, general , binding regulations and proposed changes have different legal status. Their effects also depend on the product and institution. Claims about a particular arrangement’s legality require its actual contracts and facts; the operating comparisons here do not supply that conclusion.
What would materially change the evidence
Upcoming company results can clarify whether usage growth produces retained earnings after support costs, losses and partner reimbursements. For the named examples, material developments include Chime/Stride approvals or a changed closing timetable, FinWise’s post-acquisition credit exposure and integration costs, Coastal’s partner-loss recovery and program mix, and changes in issuing/payment volume and funding costs at The Bancorp. None of these is presumed resolved before a dated disclosure.
For operational resilience, new enforcement modifications or terminations, documented customer restitution, material partner migrations, usable customer-level reconciliation evidence and changes in recordkeeping law would be meaningful updates. The November 16 comment deadline is a scheduled regulatory milestone, not a forecast of finalization.
The present evidence supports a mixed conclusion: sponsorship is operating at substantial disclosed scale and can create valuable products and fee streams, while retained economics, counterparties and continuity remain program-specific. The unresolved question is not whether distribution exists. It is how consistently that distribution delivers durable customer access and returns after the full cost of keeping the service working.
Sources
- Banking agencies, Joint Statement on Banks’ Arrangements with Third Parties to Deliver Bank Deposit Products and Services; July 25, 2024Official sourceBack to text: ↑
- Federal Reserve SR 23-4, Interagency Guidance on Third-Party Relationships: Risk Management; June 7, 2023Official sourceBack to text: ↑
- Banking agencies and NCUA, proposed third-party guidance announcement; September 11, 2026, updated September 15Official releaseBack to text: ↑
- Federal Register, Proposed Third-Party Risk Management Guidance; September 15, 2026; comments due November 16, 2026Official sourceBack to text: ↑
- FDIC Consumer News, Banking With Third-Party Apps; June 2024 issue, page updated May 31, 2024Official sourceBack to text: ↑
- Coastal Financial Corporation, Q2 2026 results; July 30, 2026; period ended June 30, 2026SourceBack to text: ↑1↑2↑3
- FinWise Bancorp, Tallied technology-platform and asset acquisition; July 20, 2026Filing / reportBack to text: ↑1↑2
- Chime, Form 8-K and Central Service Corporation merger agreement; September 8, 2026Filing / reportBack to text: ↑1↑2
- Chime, agreement to acquire Stride Bank; September 8, 2026Filing / reportBack to text: ↑1↑2
- Mercury, Column N.A. policies and banking-provider disclosures; undated, checked October 4, 2026SourceBack to text: ↑
- Pathward, expanded TabaPay agreement; August 5, 2026SourceBack to text: ↑
- Chime, Q2 2026 financial results, August 5, 2026; period ended June 30, including metric definitionsFiling / reportBack to text: ↑1↑2
- The Bancorp, Q2 2026 financial results; period ended June 30, 2026SourceBack to text: ↑
- FinWise Bancorp, Q2 2026 financial results, July 29, 2026; period ended June 30Filing / reportBack to text: ↑1↑2↑3
- Federal Reserve, Evolve enforcement announcement, June 14, 2024Official releaseBack to text: ↑
- Federal Reserve and Arkansas State Bank Department, Evolve cease-and-desist order, effective June 11, 2024, paragraphs 5, 19 and 28Official release · PDFBack to text: ↑
- FDIC, proposed third-party-account recordkeeping rule announcement, September 17, 2024; historical proposal, not presented as operative lawOfficial releaseBack to text: ↑
- OCC, order terminating Blue Ridge Bank consent order, November 13, 2025Official source · PDFBack to text: ↑
- OCC, November 2025 enforcement announcement, including Blue Ridge terminationOfficial releaseBack to text: ↑