The central thesis: access to credit depends on funding, underwriting and servicing together
Avant is a consumer-finance platform seeking to establish its own bank. Its existing branded credit products are issued by WebBank; its proposed Avant Bank, N.A. is a separate regulatory project. On September 18, 2026, the company announced applications to the OCC and FDIC. The FDIC’s list, updated September 30 with data through September 29, records the deposit-insurance application as pending. Filing is not approval, deposit insurance or authorization to open. [1] [2]
The company’s strategic problem is straightforward to state and difficult to solve: make credit available to consumers who may be expensive to serve, while collecting enough revenue to cover losses, funding, acquisition, servicing and capital. Better software can help each part of that equation. It cannot remove a borrower’s income shock, an investor’s eligibility restriction or the cost of carrying credit risk.
This article’s assessment is that Avant’s 2026 activity shows a coherent attempt to improve that equation through diversified funding, a wider customer relationship and more automated servicing. A bank charter could extend the strategy, but also add significant operating and capital obligations. The strongest evaluation focuses on realized customer and credit outcomes rather than cumulative loan volumes, an AI label or the rating of one senior securitization tranche.
Research was completed September 30, 2026. Company announcements are identified as such; rating-agency observations apply to the specified transaction and period. Current consolidated audited financials and private financing contracts were not reviewed. Calculations and scenarios below are explicitly illustrative, not estimates of Avant’s profitability.
Company background and what the scale figures actually mean
Avant began in 2012 with a focus on improving borrowing for middle-income consumers. Its current website identifies Al Goldstein as executive chairman and CEO and Paul Zhang as cofounder and chief information officer. A February 2020 announcement identifies Goldstein, John Sun and Zhang as the founders and documents the completed spin-off of Amount, the enterprise software business. Amount’s historical bank-client reach should therefore not be counted as Avant consumer-lending customers. [3] [4]
The September 15, 2026 funding announcement says Avant had served more than 4.8 million unique customers and connected them to over $17.6 billion in credit since founding. These are company-reported cumulative measures. They are not the number of active borrowers, current receivables, annual revenue or assets retained on Avant’s balance sheet. The announcement includes both cards and personal loans in the credit figure. [8]
That distinction affects every comparison. An originator can arrange a large volume of loans and sell most of them; a card business can accumulate many accounts while a smaller subset carries balances. A serious company review needs current active accounts, receivables by product, funded originations, retained exposure and repeat-customer behavior on matching dates. Historical scale is evidence of operating experience, but does not establish present earnings quality or .
Products and borrower costs: the current offer terms matter
The current primary product pages below were reviewed on September 30. Individual terms depend on underwriting, state rules and the offer channel. The loan page shows an administration fee deducted from proceeds; a borrower can therefore owe more principal than the cash received. The card disclosures distinguish direct website applications from other offers. [5] [6]
Avant’s agreement page also distinguishes introductory annual-fee cards from other cards, with different first-year and later fee schedules. Some later annual fees are billed monthly. These categories should not be combined into a claim that every card has both an annual charge and an additional monthly charge. The applicant’s actual agreement is the relevant document. [7]
For an illustrative fee calculation, a $10,000 loan with a 9.99% administration fee would deliver $9,001 before any other adjustments, while starting with $10,000 of principal. That is arithmetic using an advertised maximum fee, not a quote or a claim that every borrower pays it. Compare the amount financed, , scheduled payments and total payments together; the stated interest rate alone can conceal the effect of an upfront deduction.
Similarly, a $75 annual fee on a $300 starting card limit equals 25% of that limit. This ratio is not an APR and is not the card’s total cost. It shows why a modest dollar fee can matter greatly on a small line. Consumers who carry balances must also consider interest, while customers seeking credit-building benefits should compare the full cost and usability of available alternatives.
Scroll horizontally to see all columns.
| Product | Publicly advertised terms | Important qualification |
|---|---|---|
| Personal loan | $2,000–$35,000; 24–60 months; 9.95%–35.99% APR; administration fee up to 9.99%. | Current general loan page. State minimums and the actual offer may differ. [5] |
| Avant credit card | $300–$3,000 advertised credit limits; general APR range 29.99%–35.99%. | Eligibility and offer channel determine the actual terms. [6] |
| Direct website card offer | $0–$75 annual membership fee and 35.99% APR in the current disclosure. | Mail and other offers may have different pricing. [6] |
| Credit Builder | Free monitoring and financial-health features offered to existing customers through the app. | The announced product is a financial-health tool, rather than a new installment loan. [13] |
The operating model: the brand, lender, asset owner and servicer can differ
Avant’s product disclosures identify WebBank as issuer. Fitch’s February 2026 transaction analysis describes WebBank-originated loans passing through Avant-related entities into the securitization, with Avant servicing the receivables. This is a description of that transaction, not a universal ownership map for every Avant account. [5] [12]
Four roles matter. The originator makes the legal extension of credit. The owner bears the economic exposure defined by its asset and contractual rights. The servicer administers payments, customer communications and collections. The funding provider supplies capital against specified assets or purchases eligible loans. A single brand visible to a borrower does not establish that one company performs all four roles.
For Avant, this structure can distribute financing across several channels while keeping the customer experience relatively consistent. For a bank or capital partner, it makes reconciliation and oversight essential. The account agreement, servicing system, investor ledger and cash movements must describe the same obligation even when the asset changes hands.
The key diligence questions are where credit losses ultimately sit, what exposure Avant retains, what representations or repurchase obligations survive a sale, and what happens when a loan fails an investor’s eligibility criteria. Public announcements do not disclose all those terms. Calling a model capital-light without inspecting the retained interests, guarantees and pipeline would be premature.
Three 2026 transactions illustrate different kinds of capital
Avant’s funding announcements show several mechanisms rather than one interchangeable pool of money. The February securitization recycles collections during a revolving period. The July facility provides asset-based financing, with Avant’s CEO pointing to card-portfolio growth. The September agreement creates a conditional channel for future personal-loan sales. [8] [9] [10]
These headline amounts should not be added and presented as unrestricted cash, new equity or incremental outstanding assets. They refer to different products, time horizons and forms of financing. A revolving transaction can finance more cumulative production than its initial note balance, while a forward-flow commitment can remain partly unused if origination volume or eligibility falls short.
The diversification benefit is plausible: one channel can remain available when another becomes less attractive. It is incomplete protection because capital providers can respond to the same deterioration in consumer credit. The relevant stress test asks what happens when losses rise, fall and whole-loan buyers tighten criteria at the same time.
Scroll horizontally to see all columns.
| Announcement | Documented transaction | What it establishes |
|---|---|---|
| February 26, 2026 | $200 million personal-loan securitization; company-described 24-month revolving period and expected lifetime financing capacity above $500 million. [10] | Access to structured funding. Expected recycling is not additional equity or a guaranteed volume outcome. |
| July 29, 2026 | Aperture Investors announced a $300 million asset-based financing facility with Avant. [9] | An additional financing relationship. Public terms do not establish fully drawn cash or the complete package. |
| September 15, 2026 | InterVest affiliate committed to purchase up to $1 billion of personal loans over 24 months, subject to customary funding and financing conditions. [8] | A conditional forward-flow channel. The maximum commitment is not a completed sale of $1 billion of loans. |
AAA notes do not make the underlying borrowers—or Avant—AAA
KBRA’s February 20 report identifies six note classes totaling $200 million in Avant Loans Funding Trust 2026-REV1. It reports initial hard credit enhancement of 67.38% for Class A and 3.24% for Class F, with structural protection including subordination and overcollateralization. Fitch assigned Class A an AAAsf rating and progressively lower ratings down the structure. These are security-specific assessments, not an Avant corporate credit rating. [11] [12]
The mechanism is loss allocation. Junior interests, excess asset value and other protections can absorb losses before a senior investor is affected. A pool of materially risky loans can therefore support a highly rated senior class. Someone still bears the losses and the cost of providing that protection; an attractive senior funding rate alone does not reveal the residual owner’s return.
A revolving pool adds a second issue: future collateral can differ from the loans observed at closing. Eligibility rules, concentration limits and performance triggers constrain that change. When examining a transaction, ask how cash is allocated after a trigger breach and whether the structure can stop replenishing assets before losses materially erode protection.
The company release and rating agencies use different historical transaction counts. This profile does not combine those counts into a single track-record claim. The relevant verified event is the named February 2026 transaction and its stated structure. [10] [11]
Credit evidence is useful when its denominator and date stay attached
Fitch’s February analysis reports a weighted-average FICO score of 646 and weighted-average of 25.31% for the specified pool; all loans were current at its December 15, 2025 statistical cutoff. It describes weaker 2021–2022 origination performance and subsequent improvement that remained above 2017–2020 default levels. Its 19.14% gross-default assumption applies to a modeled worst-case revolving portfolio. It is not a current company-wide rate, annual rate or realized loss. [12]
Those observations support a more precise description than either a prime-loan story or an undifferentiated subprime label. The pool’s borrowers had varied credit characteristics, and the securitized selection is not necessarily representative of the whole platform. Loan pricing reflects several costs and risks; it cannot by itself establish how well a model ranks applicants or whether the product is affordable.
The best ongoing evidence would be curves by origination quarter, product, score band, channel and term. Compare cumulative defaults and recoveries at the same months on book. A rapidly growing portfolio can look healthier because new loans have had less time to default; a shrinking portfolio can look worse as remaining loans season. Average delinquency without that context can mislead.
For cards, separate activation, utilization, payment rates, line increases and loss emergence. For installment loans, track prepayments and repeat borrowing as well as defaults. A borrower who repays by taking another expensive loan creates a different customer-outcome question from a borrower whose income comfortably supports repayment. Those distinctions are analytical priorities, not observed conclusions about Avant’s current borrowers.
A simple earnings bridge shows why funding and losses dominate
Consider a hypothetical portfolio with $100 million of average receivables and an annual realized revenue yield of 25%. Assume $12 million of net credit losses, $7 million of funding expense and $4 million of acquisition and servicing expense. The remaining contribution is $2 million before central overhead, taxes and capital costs. None of these inputs is an estimate of Avant’s financial statements.
If credit losses rise by three percentage points, the same portfolio produces a $1 million shortfall. If funding costs alone fall by two percentage points, contribution rises to $4 million. Combining both changes produces $1 million. A lower cost of funds helps, but a deterioration in borrower performance can consume that benefit quickly.
This simplified bridge deliberately avoids subtracting a lifetime default assumption from an annual yield. That would mix periods and gross defaults with net losses. An actual model also needs amortization, recoveries, fee accounting, loan-sale gains or discounts, retained interests and capital consumption. Off-balance-sheet originations require a separate fee-and-servicing model rather than being treated as owned receivables.
Customer acquisition is another constraint. A profitable existing account can become an unprofitable new cohort if marketing and partner costs rise faster than expected lifetime contribution. Evaluate return by channel and after credit losses, rather than relying on a blended cost per application or a company-wide approval rate.
Scroll horizontally to see all columns.
| Illustrative annual scenario | Contribution on $100 million average receivables |
|---|---|
| Base: 25% revenue yield − 12% losses − 7% funding − 4% operating costs | $2 million |
| Credit losses rise to 15%; other inputs unchanged | −$1 million |
| Funding expense falls to 5%; base losses unchanged | $4 million |
| Both higher losses and lower funding expense | $1 million |
The proposed bank: a funding opportunity with substantial execution requirements
Avant’s September application announcement names Charles Whittaker as proposed bank CEO, Kevin Friedrich as proposed CFO and Margaret Hermes as proposed president and COO. Management argues that deposits and more direct control of products could lower funding costs and simplify operations. The regulatory status verified here is the FDIC’s pending application listing; no approval or opening date was established in this review. [1] [2]
The economic appeal is a closer relationship between customer acquisition, deposits and lending. A stable deposit base could complement institutional funding and reduce reliance on repeated asset sales. But the relevant comparison is the fully loaded cost of deposits: interest, marketing, technology, customer service, , insurance and compliance. A headline deposit rate is not the all-in cost of funding.
A banking model also requires a credible balance-sheet strategy. Consumer assets and deposits can respond differently to stress. Customers may move money quickly even while installment-loan cash flows arrive slowly. A business plan needs to demonstrate how liquidity, capital and contingency funding work together, including a scenario in which wholesale funding becomes less available.
Approval, if obtained, would begin a transition rather than instantly replace existing contracts and operations. Questions include which products and assets move into the bank, what remains outside it, how customer agreements change, and how capital is allocated between entities. A successful outcome would be a more resilient and efficient institution; a weak outcome would add fixed costs without enough durable funding or customer benefit.
AI servicing: a documented deployment, with outcome questions still open
Avant’s May 5, 2026 announcement identifies Replicant as the partner for a virtual agent across Avant Card inbound customer service, launched in January. The company reported that approximately 62% of calls were completed without transfer to a human agent and that post-call survey respondents rated it 4.6 out of 5. These are attributed operating claims from the announcement, not independently audited results or a September measurement. [14]
Call containment is useful but incomplete. A call can finish without a transfer even when the customer later calls back, abandons a disputed issue or receives an incorrect explanation. Measure repeat contacts, correct resolution, complaint escalation, payment authorization errors and time to a capable human. Survey results need response rates and sampling details before they can represent the whole customer population.
The high-value use cases are narrow and testable: retrieving account information, explaining a documented process and completing a clearly authorized routine task. More consequential actions need tighter controls, including disputed amounts, hardship decisions, payment-plan changes and contested payoff figures. Preserve the underlying record and the authority for any action so an institution can reconstruct what happened.
There is a historical connection to Spring Labs through John Sun, who helped found Avant and is listed as Spring Labs’ current CEO. That does not establish common ownership or a specific production integration. The virtual-agent deployment documented here names Replicant; the entrepreneurial connection should not be used to substitute a different vendor. [4] [14] [18]
Credit Builder and co-branding aim to make the relationship broader
Avant’s January 28 release describes Credit Builder, developed with SavvyMoney, as a free app-based financial-health tool for existing customers. The company reported 357,000 users by December 2025 following an April pilot, and said 75% had steady or improved scores since enrollment. Combining unchanged and improved scores is important: this is not a claim that 75% improved, nor proof that the tool caused the outcome. [13]
The potential business value is engagement between borrowing events. Customers who understand their balances and progress may return more often and need fewer basic explanations. Evaluating the tool’s effect requires a comparable control group and attention to self-selection: customers who voluntarily monitor credit may already behave differently from those who do not.
NASCAR’s August 13 announcement describes an exclusive U.S. co-branded credit-card partnership beginning January 1, 2027, with launch expected that month. Announced benefits include integration with the fan-rewards program and enhanced rewards. Product terms and eligibility were to be announced nearer launch. This is a planned distribution expansion, not a live portfolio with demonstrated economics. [15]
Co-branding can provide access to an engaged audience and increase reasons to use a card. It also introduces partner payments, rewards costs and concentration in a particular acquisition channel. The meaningful test is incremental risk-adjusted contribution and retention after those costs, rather than applications generated by a recognizable brand.
Servicing history makes payment controls especially important
The FTC sued Avant in April 2019 over alleged unfair and deceptive servicing practices, including unauthorized charges, misleading payment-method representations and payoff problems. The signed stipulated order imposed a $3.85 million monetary judgment and conduct restrictions. In January 2022, the FTC announced more than $3.7 million in refunds to 17,367 consumers. These are historical proceedings and redress, not a newly announced 2026 enforcement action. [16] [17]
The order addresses misrepresentations, payment handling and charges without express informed consent. Its detailed payment-posting and payoff provisions in Section II expressly exclude certain open-end credit-card accounts subject to Regulation Z; the order should not be paraphrased as though every subsection applies identically to every product. The signed text matters more than a broad summary of the case. [16]
The analytical lesson is that servicing accuracy is part of both customer protection and credit economics. Incorrect payment posting can create false , inappropriate collections and inaccurate records. A technically impressive underwriting process cannot compensate for an unreliable ledger or unauthorized debit. Historical issues warrant targeted evidence of current controls, without assuming that the earlier allegations describe present conduct.
For a bank or financing partner, useful tests would trace a payment from receipt to ledger and customer statement, reproduce a payoff quote, check authorization for a changed debit, and confirm how a disputed transaction reaches a human owner. Apply those tests to automated and human channels alike. Complaint counts and customer reviews can identify leads, but without comparable denominators and case verification they do not establish a company-wide misconduct rate.
Competitive position: several plausible strengths, several unresolved questions
Avant’s potential strengths are accumulated consumer-credit operating experience, a combination of loans and cards, documented access to several funding channels and an existing customer relationship that can support additional services. The challenge is turning those capabilities into better outcomes at an acceptable cost through a full credit cycle. Each strength has a corresponding test: performance, contribution by product, usable and retention after acquisition costs.
Consumers can compare its products with bank and credit-union loans, other digital lenders, secured cards and unsecured card issuers. Investors and partners should compare lenders serving similar borrowers, terms and channels. Comparing Avant’s losses with a prime lender without adjusting the borrower mix says little about underwriting quality. Comparing alone omits fees, term, access and repayment affordability.
Management’s references to proprietary decisioning and machine learning describe the company’s approach, but public materials reviewed do not establish a reproducible advantage over alternatives on matched applicants. Evidence would include performance at equal approval rates, stability after model changes, reliable explanations and outcomes for relevant customer groups. More complex models are useful when they improve those results and remain governable.
The principal unresolved company-level questions are current revenue and profitability, liquidity and debt maturities, retained credit exposure, funding , borrower-level affordability outcomes and the capital required for the proposed bank. Public transaction announcements answer pieces of that picture. They do not add up to a consolidated balance sheet or a valuation.
What to watch next
The clearest near-term milestones are official action on the bank application, actual execution under the forward-flow agreement, updated transaction surveillance and the planned January 2027 NASCAR card launch. For servicing technology, the next useful evidence would connect automation to durable resolution, repeat-contact rates and error reduction, rather than simply reporting a higher share of calls without transfer.
The constructive case would strengthen with stable mature- losses, demonstrable funding diversification during stress, repeat borrowing that improves customer finances, and a bank plan supported by sufficient capital and operational readiness. The case would weaken if growth depends on looser underwriting, increasing customer cost or financing terms that become restrictive just when losses rise.
Avant merits attention because the company brings several important consumer-finance questions into one business: how to price and fund non-prime credit, how to preserve customer trust at scale, and whether a digital lender gains enough from a bank charter to justify the added responsibilities. The answer will be found in cohort outcomes, cash flows and execution.
Sources
- Avant — application announcement, proposed leadership and strategy, September 18, 2026SourceBack to text: ↑1↑2
- FDIC — pending new deposit-insurance applications; data through September 29, 2026, page updated September 30Official sourceBack to text: ↑1↑2
- Avant — current company history and leadership page, reviewed September 30, 2026SourceBack to text: ↑
- Amount / Avant — completed Amount spin-off and founding history, February 5, 2020SourceBack to text: ↑1↑2
- Avant — current personal-loan product, fee and issuer disclosuresSourceBack to text: ↑1↑2↑3↑4
- Avant — current credit-card product and offer-channel disclosuresSourceBack to text: ↑1↑2↑3↑4
- Avant — cardmember-agreement categories and fee schedulesSourceBack to text: ↑
- Avant and InterVest — conditional $1 billion forward-flow agreement and dated cumulative scale, September 15, 2026SourceBack to text: ↑1↑2↑3
- Aperture Investors — $300 million asset-based financing facility, July 29, 2026SourceBack to text: ↑1↑2
- Avant — $200 million personal-loan securitization announcement, February 26, 2026SourceBack to text: ↑1↑2↑3
- KBRA — Avant Loans Funding Trust 2026-REV1 ratings and structural enhancement, February 20, 2026SourceBack to text: ↑1↑2
- Fitch — AVNT 2026-REV1 final ratings, collateral, historical trends and modeled defaults, February 20, 2026SourceBack to text: ↑1↑2↑3
- Avant — Credit Builder with SavvyMoney and attributed pilot results, January 28, 2026SourceBack to text: ↑1↑2
- Avant — Replicant virtual-agent deployment and company-reported results, May 5, 2026SourceBack to text: ↑1↑2
- NASCAR — official co-branded card announcement and planned January 2027 launch, August 13, 2026SourceBack to text: ↑
- U.S. District Court / FTC — signed stipulated injunction and monetary judgment, May 2019Official source · PDFBack to text: ↑1↑2
- FTC — historical allegations and consumer refunds, January 25, 2022Official releaseBack to text: ↑
- Spring Labs — current leadership page, reviewed September 30, 2026SourceBack to text: ↑