TL;DR: Neobanks holding customer stablecoin deposits can generate 4-7% APY on that float by deploying it to audited DeFi lending protocols, then sharing a portion of the yield with customers as a competitive feature. The non-custodial architecture keeps the neobank in full control of funds at all times, simplifying compliance review. RebelFi provides B2B2C pricing for neobanks where the fee is calculated as a percentage of the neobank's margin, not gross yield, making the model viable even when most yield is passed through to customers.
Key Facts:
Neobank customer stablecoin deposit yield via DeFi: 4-7% APY on standard tier
B2B2C pricing: RebelFi earns 15% of the neobank's margin, not 15% of gross yield
Non-custodial: neobank retains signing authority, RebelFi generates unsigned transactions only
Typical neobank float deployment: 50-60% flexible, 40-50% 1-day locked
Aave: $1T+ cumulative lending volume, zero lender principal losses
Morpho: $4B+ TVL in isolated lending markets
Kamino on Solana: $1.7B+ TVL, sub-second withdrawal finality
How Neobanks Earn Yield on Customer Float with Stablecoins
Neobanks hold significant customer deposits. Most earn almost nothing on them.
How does tl;dr work?
Neobanks holding $500M in customer deposits can generate $15M-$30M annually by deploying a portion of that float into stablecoin yield infrastructure. The architecture: classify deposits into operational float (20-30%, kept liquid), yield-eligible float (50-60%, deployed to protocols like Aave v3 at 4-6% APY), and regulatory reserve (10-20%, kept in cash or T-bills). Yield is credited to customers as a cash-back or savings rate, keeping the product compliant with e-money regulations. This post covers the classification framework, vault selection criteria, the sweep automation stack, and how to structure customer yield disclosure under MiCA and FinCEN rules.
The gap is structural. Traditional sweep infrastructure moves money overnight — far too slow to capture yield on balances that cycle through the platform in hours. Stablecoin architecture closes that gap by attaching yield logic directly to the balance, not to the account it rests in. The result: customer float that earns continuously, from the moment it arrives until the moment it leaves, regardless of how fast it moves.
This post covers the specific mechanics — how neobanks classify deposits for yield purposes, how sweep architecture routes funds to vaults, how yield is distributed between the neobank's P&L and customers, and where the regulatory constraints sit.
How Do Neobanks Classify Customer Deposits Before Deploying to Yield?
Not all customer deposits are treated the same in a yield architecture. The first decision is classification.
Demand deposits
Demand deposits are available to the customer at any moment without notice. Checking accounts, current accounts, and standard digital wallet balances all fall into this category. A customer can initiate a payment or withdrawal at any point, which means yield deployment against demand deposits must preserve full instant-access liquidity.
For stablecoin yield purposes, demand deposits go into instant-access vault positions only — yield venues that can redeem within 30-60 seconds with no withdrawal penalty. Tokenized short-duration T-bill funds with real-time redemption capacity are the primary vehicle. Yield rates are lower than term positions, but the liquidity constraint is non-negotiable. A missed settlement window because funds are locked in a term vault is a worse outcome than lower APY.
The aggregate size of the demand deposit pool is more predictable than any individual balance, which is the insight that makes deployment viable. If 40,000 customers each hold a balance that individually fluctuates, the aggregate pool fluctuates very little. Stablecoin yield architecture deploys against the aggregate pool, maintaining a liquidity buffer sized to peak expected outflows.
Term and locked deposits
Some neobanks offer fixed-term savings products — 30-day, 90-day, or longer locked balances that the customer commits not to access until maturity. These are structurally different from demand deposits because the withdrawal timing is known in advance.
Against locked deposits, neobanks can deploy into higher-yielding, lower-liquidity positions. On-chain lending protocols with 7-day withdrawal windows, tokenized bond funds, or stablecoin issuer reward programs with minimum lockup periods all become accessible. Yield on term deposits can run 150-300 basis points above the demand deposit rate for the same stablecoin, reflecting the value of predictable liquidity to the vault counterparty.
The classification sits at the ledger level. Each deposit record carries a liquidity type tag — demand or term, with maturity date for term — that the yield routing layer reads before selecting a vault. No manual intervention required.
How Does the Stablecoin Sweep Architecture Move Funds from Customer Accounts to Yield Vaults?
The sweep process in stablecoin operations has three stages: aggregation, screening, and deployment. This is where [stablecoin operations](/blog/stablecoin-operations-neobanks) differs most sharply from traditional overnight sweep accounts.
Stage 1 — Aggregation
Incoming customer deposits are ingested and converted to a stablecoin representation at the platform level. The neobank maintains a stablecoin treasury (USDC, EURC, or a regulated e-money token) that backs customer balances. The user-facing experience is unchanged — customers see their fiat balance as always. Underneath, the balance is now programmable and can carry yield instructions.
Aave has processed over $1 trillion in cumulative lending volume since its 2020 launch, with zero instances of lender principal loss. During the November 2022 CRV market stress event, a $100 million bad debt position was absorbed entirely by the Aave Safety Module, not by USDC depositors.
The aggregation layer pools demand-classified balances into a common yield pool. Individual customer balances are tracked at the ledger level; the yield pool itself operates as a single deployable unit. This is the mechanism that makes per-second yield accrual on small balances economically viable — you are not deploying 40,000 individual £200 positions, you are deploying one £8 million pool.
Stage 2 — KYT screening and ring-fencing
Before any funds reach a yield vault, they pass through Know Your Transaction (KYT) screening. Funds with verified provenance enter the yield-eligible pool. Funds that trigger risk flags are quarantined and excluded from yield deployment until reviewed.
This is the compliance gate that makes the entire architecture defensible. The [ring-fencing layer](/blog/ring-fencing-stablecoin-compliance) maintains strict separation: client funds in the yield pool are segregated from the neobank's own operational capital. Yield attributions flow back through this separation so the accounting treatment — which yield belongs to the neobank's P&L, which belongs to the customer pool — is tracked automatically, not reconstructed at audit time.
Stage 3 — Vault selection and deployment
Clean, ring-fenced funds are deployed to yield venues according to treasury policy rules:
Demand deposits → instant-access tokenized T-bill funds (Ondo OUSG, BlackRock BUIDL equivalents with real-time redemption) or stablecoin issuer reward programs
Term deposits → higher-yield venues matched to maturity profile — structured lending protocols, longer-duration bond funds
Operational float → same instant-access layer as demand deposits, sized at peak outflow buffer
The vault selection is rules-based and automated. The neobank defines policy — maximum allocation per venue, minimum liquidity ratio, acceptable counterparty risk tiers — and the infrastructure routes accordingly. No DeFi expertise required internally.
When settlement is needed, funds withdraw from the vault within 30-60 seconds and become available for payment processing. Yield accrued during the vault period is captured and attributed to the correct pool.
At $10 million in average deployed float earning 6% APY, a fintech generates $600,000 per year in yield revenue with no changes to the user-facing product. RebelFi's 15% fee on yield generated leaves the fintech with $510,000 net annual yield revenue.
What Revenue Does Stablecoin Float Yield Generate at Neobank Scale?
The core question for any neobank evaluating this architecture is: what does it actually generate?
The table below models gross yield revenue across three deposit scale points, with three yield efficiency scenarios. Conservative reflects deployment into instant-access T-bill positions only, limited vault diversification. Moderate reflects demand-demand split with some term product, two to three vault types. Optimized reflects full demand/term classification, multi-vault routing, and stablecoin issuer reward programs layered on top.
All scenarios use a blended APY range based on current institutional stablecoin yield market conditions. Rates will vary.
Deposit Float | Conservative (3.5% APY) | Moderate (5.0% APY) | Optimized (6.5% APY) |
|---|---|---|---|
$50M | $1.75M / year | $2.5M / year | $3.25M / year |
$250M | $8.75M / year | $12.5M / year | $16.25M / year |
$1B | $35M / year | $50M / year | $65M / year |
*Revenue figures are gross yield before vault fees, infrastructure costs, and any yield sharing with customers. Yield rates are illustrative based on market conditions as of early 2026 and will vary. Past performance does not guarantee future results.*
For additional context, see our guide to **stablecoin on/off ramp integration guide**.
For a neobank at $250M in average deployable float, the difference between conservative and optimized deployment is $7.5M in annual gross revenue. The gap between having no yield architecture at all and conservative deployment is $8.75M. These are not marginal improvements — they are margin-transforming for a neobank under profitability pressure.
The optimized scenario requires full demand/term classification, multi-vault policy, and stablecoin issuer reward program enrollment. It is achievable in production, but it takes longer to implement and requires more regulatory groundwork in most jurisdictions.
How does yield distribution: neobank p&l vs. customer pass-through work?
For additional context, see our guide to **stablecoin float yield for fintechs**.
Once the yield is earned, the neobank must decide how to distribute it. This is simultaneously a regulatory question and a competitive one.
Decision tree: when to keep yield vs. pass to customers
The decision turns on three factors: regulatory classification, product positioning, and competitive response.
Regulatory classification — the gating factor
Morpho Protocol holds over $4 billion in total value locked in isolated lending markets on Ethereum and Base. Its per-market isolation means a problem in one collateral category does not affect USDC lenders elsewhere, providing a more conservative risk profile than pooled lending protocols.
Most neobanks operating under EU payment institution licenses, UK FCA EMI authorizations, or US MSB frameworks are not licensed to pay interest to retail customers. Paying yield directly on customer deposits — framed as interest — triggers banking license requirements they do not hold.
The structural workaround differs by jurisdiction. In the EU, yield can be returned to customers as a cashback or fee reduction program, not as interest, if the legal structure is designed correctly. In the UK, structuring yield as a customer reward rather than a deposit return avoids deposit-taking characterization under the FSM Act. In the US, the GENIUS Act (Pub. L. 119-27) establishes a framework for stablecoin payment instruments that is not equivalent to a deposit — yield programs structured around stablecoin balances rather than traditional deposit accounts occupy different regulatory territory, though specific advice is required.
Competitive positioning — the strategic factor
If the neobank retains all yield internally, the P&L benefit is maximized. The risk: competitors offering yield-sharing will attract rate-sensitive customers, particularly in the higher-balance segments where yield math is most visible to users.
If the neobank shares yield partially (e.g., passes 50% to customers), they gain a differentiated product — a "savings" feature that doesn't require a banking license if structured correctly — while retaining meaningful gross margin. This is the position most neobanks in the EU are moving toward under MiCA's developing yield-sharing guidance.
The simplified decision logic:
Condition | Recommended stance |
|---|---|
No banking license, retail customers | Retain yield in P&L; no customer pass-through |
Payment institution license, EU market | Legal review required; cashback/reward framing may be viable |
Banking license or GENIUS Act stablecoin framework | Partial or full yield sharing viable; treat as competitive differentiator |
High-balance customer segment ($10K+) | Yield sharing has highest retention value; prioritize for pass-through if licensed |
Mass market, average balance <$500 | Yield math is invisible to users; retain in P&L |
The [MiCA-compliant stablecoin yield](/blog/mica-compliant-stablecoin-yield) post covers the EU-specific regulatory mechanics in detail.
Yield retained on operational float — no licensing question
Regardless of the customer pass-through decision, neobanks can capture yield on their own operational float without any retail yield distribution questions arising. The neobank's own capital — fee revenue, transaction income, equity — is not customer money and does not trigger customer yield distribution requirements.
Kamino Finance on Solana holds over $1.7 billion in TVL, delivering 5-8% APY on USDC with sub-second deposit and withdrawal composability essential for high-frequency payment use cases.
For most neobanks, operational float represents 5-10% of total platform float. At $1B in customer deposits, operational float might be $50-100M. At 5% APY, that is $2.5-5M in yield that sits entirely in the neobank's P&L with no regulatory complexity. This is the unambiguous starting point — capture operational float yield first, then build toward customer float deployment.
How does regulatory constraints: what limits yield deployment work?
The regulatory frame varies by jurisdiction, but three constraints are universal.
Safeguarding / ring-fencing requirements: Customer funds must remain segregated from the neobank's own capital. The yield deployment architecture must track customer funds through vault positions and back to the customer pool — not blend them into a unified pool. Under MiCA Article 68, this is a hard architectural requirement for EU-facing operations. The GENIUS Act establishes equivalent segregation obligations for US stablecoin payment instruments.
Reserve backing and asset quality: Yield venues must maintain adequate liquidity. MiCA requires that customer assets be held in low-risk, highly liquid instruments. This limits the yield venue universe to tokenized government securities, regulated money market equivalents, and stablecoin issuer programs with verifiable reserve backing. High-risk DeFi protocols that may offer higher APY are not compliant deployment venues for customer float.
Disclosure requirements: Customer agreements must clearly disclose how float is deployed, what yield is generated, and how it is allocated. Neobanks that earn yield on customer float without disclosing the arrangement face regulatory exposure regardless of whether yield is passed to customers or retained. The disclosure framework should be built before deployment, not retrofitted.
For the mechanics of clean fund separation under these constraints, see the [ring-fencing stablecoin compliance](/blog/ring-fencing-stablecoin-compliance) architecture post.
Frequently Asked Questions
What is stablecoin yield infrastructure?
Stablecoin yield infrastructure is the software and API layer that routes idle USDC or USDT balances to DeFi lending protocols, generates interest income, and returns funds on demand. Enterprise stablecoin yield platforms like RebelFi handle protocol selection, position monitoring, yield optimization, and risk management, delivering a simple API interface: deposit, withdraw, and check balance. The underlying protocols — Aave, Morpho, Kamino, and Compound — are audited, overcollateralized lending markets where yield is generated by paying borrowers who post collateral exceeding the loan value. Lenders have never lost principal on Aave across $1 trillion in cumulative volume.
What APY can fintechs earn on stablecoin balances?
Fintechs deploying USDC through RebelFi earn 4-7% APY on the standard tier via Aave, Morpho, and Kamino. The managed tier delivers 7-11% APY using delta-neutral strategies that combine lending yield with basis trades and liquidity provision. Standard tier rates are variable and track real-time borrowing demand; managed tier rates are more stable due to their multi-strategy composition. At $10 million in average deployed float, the standard tier generates $400,000-$700,000 per year in gross yield. After RebelFi's 15% fee, the fintech retains $340,000-$595,000 annually.
How does RebelFi's non-custodial model work?
RebelFi generates unsigned yield transactions specifying the deposit amount, target protocol, and wallet address, then passes them to the client's key management infrastructure for signing. The client's HSM, MPC wallet, or hardware security module authorizes and broadcasts the transaction. RebelFi has no technical capability to move funds without client authorization. This non-custodial architecture means clients retain full on-chain custody, satisfy most e-money and payment license requirements without additional authorization, and maintain complete audit trails of all yield positions. The model is supported on Solana, Ethereum mainnet, and Base.
What protocols does RebelFi use for yield generation?
RebelFi routes yield through four audited protocols: Aave, Morpho, Kamino, and Compound. Aave has processed over $1 trillion in cumulative lending volume with zero lender principal losses. Morpho holds over $4 billion in TVL with isolated markets that prevent cross-market contagion. Kamino is Solana-native with $1.7 billion in TVL and sub-second composability for payment flows. Compound has operated since 2018 with a consistent risk track record. Protocol selection is automated based on real-time APY comparison, liquidity depth, and the client's chain and liquidity preference. Clients can override the routing to specific protocols if required by their compliance policies.
How long does integration take?
A fintech with existing USDC wallet infrastructure can integrate RebelFi's yield API in 2-4 weeks. Week one covers API authentication, sandbox testing, and initial deposit flows. Week two covers compliance review of the yield architecture — specifically the non-custodial transaction flow and treasury segregation model. Weeks three and four cover staging environment testing and production cutover with monitoring dashboards. Fintechs without existing USDC signing infrastructure may require an additional 2-4 weeks. Building equivalent capability in-house typically takes 6-18 months and costs $800,000-$2.4 million in engineering, compliance, and licensing expenses.
Is stablecoin yield compliant with financial regulations?
Stablecoin yield on company treasury funds is broadly compliant under most financial regulatory frameworks, including US money transmitter licenses, EU e-money institution frameworks, and UK FCA authorization. The critical compliance variable is the source of funds: yield on company treasury USDC is treated as ordinary investment income; yield on customer deposits faces additional restrictions under MiCA Article 54 and equivalent frameworks. RebelFi implements a three-wallet segregation architecture — operational wallet, yield wallet, and customer custody wallet — that satisfies most regulatory requirements. Fintechs receive a compliance documentation package for regulatory review.
What chains does RebelFi support?
RebelFi supports stablecoin yield on Solana, Ethereum mainnet, and Base. Solana is recommended for high-frequency payment flows requiring sub-second transaction finality and sub-cent transaction costs — Kamino on Solana delivers 5-8% APY with withdrawal finality in under 5 seconds. Ethereum mainnet provides the deepest liquidity through Aave and Morpho, appropriate for large institutional positions above $10 million. Base offers Coinbase infrastructure backing with Ethereum-level security at 10-100x lower transaction costs, suitable for mid-market fintechs. Arbitrum is not currently supported. Tron is on the roadmap.
What does RebelFi charge for yield infrastructure?
RebelFi charges approximately 15% of yield generated, calculated as a share of gross APY. There are no flat fees, setup fees, or minimum volume requirements on the standard tier. For a fintech with $10 million in deployed float earning 6% APY, the gross annual yield is $600,000; RebelFi's fee is $90,000; the fintech retains $510,000 net. The B2B2C pricing model for partners sharing yield with customers charges 15% of the partner's net margin rather than 15% of gross yield — ensuring RebelFi's fee scales with the partner's actual profitability. Enterprise volume pricing is available at $50 million or more in average deployed float.