Stablecoin volume in 2025 surpassed Visa’s transaction volume. Most of that is on-chain trading, not real-economy payments — but the share that’s real-economy payments is growing, and the part growing fastest is B2B cross-border. The reason is simple: a stablecoin transfer costs cents and settles in seconds; a wire transfer costs $25–60 and settles in 1–3 business days.

We’ve built B2B payment products on stablecoin rails for clients in commerce, payroll, and treasury. This is the architecture and the operational reality of doing it well.

The basic flow

A B2B stablecoin payment product, abstracted, looks like this:

  1. Sender’s funds enter the system in some currency (USD via ACH, USDC on-chain, etc.).
  2. If conversion is needed, an off-ramp/on-ramp partner handles fiat ↔ stablecoin.
  3. The stablecoin moves from sender’s wallet to receiver’s wallet on-chain.
  4. If conversion is needed on the other end, an on-ramp/off-ramp partner converts back.
  5. Receiver gets credited in their preferred currency.

The simplest product (USDC → USDC, both parties already on-chain) is dramatically simpler than what most B2B users actually want, which is USD → USD with a stablecoin pipe in the middle. The complexity isn’t in the on-chain transfer; it’s in the off-ramp/on-ramp experience and the operational machinery around it.

The “boring” parts that decide whether you have a product

What kills B2B payment products isn’t the smart contract layer. It’s compliance, ops, and money flow management. We’re going to skip the smart contract architecture (it’s mostly boring — ERC-20 transfers, optionally with permit-based gasless flows) and spend the article on the parts that actually decide whether the product works.

Treasury and float management

Your product needs to hold inventory of multiple assets to provide good user experience. If a user wants to send $50,000 from their USD bank account to a USDC wallet, you can’t wait for ACH (1–3 days) to clear before crediting them; you front the USDC from your treasury and wait for the ACH to settle into your bank.

This means your treasury is taking the credit risk and the timing risk. If the ACH gets reversed (returns are common in the first 60 days), you’ve sent USDC against funds you don’t have. Mitigations:

  • Tier user limits by KYC level and history. New users get tight limits; established users get larger ones.
  • Hold ACH for the float period before fronting funds for new users.
  • Maintain a reserve fund sized to cover expected returns at your scale.

The economics: float income (interest on user balances held during settlement) and FX spread are the main revenue drivers in this product. Margins are thinner than traditional payments, but volumes can be much higher because the unit costs are lower.

Compliance and KYC

If you’re moving money on behalf of users, you’re either a money transmitter or you’re working with one. There’s no third option. The cost of compliance is real:

  • US money transmitter licenses cost $1–3M to acquire across all states (or use a partner like Synapse, Bridge, Circle’s Mint).
  • Ongoing compliance staffing is $500k–$1M+ per year for a serious operation.
  • KYC tooling (Persona, Alloy, Sumsub) is $0.50–$5 per verification.
  • Sanctions screening, transaction monitoring, suspicious activity reporting — all real ongoing costs.

Most early-stage products partner with a regulated entity (Circle Mint, Bridge, Stripe’s Crypto APIs) rather than acquiring licenses themselves. The economics: pay 0.5–1% of volume to the partner; ship in months instead of years; avoid the regulatory risk of being unlicensed.

The decision matrix: if you’re under $100M annual volume and B2B (not consumer), partnering is almost always right. If you’re over $1B annual volume, you’ll likely want to acquire licenses for cost reasons. Between those two, it depends.

Off-ramp and on-ramp partners

The user experience of “send USD, receive USD with stablecoins in the middle” depends on having reliable, fast off-ramp and on-ramp partners on both ends. The market for these partners is real but uneven by jurisdiction:

  • US: Circle, Bridge, Coinbase Prime, Stripe Crypto.
  • EU: Circle, Striga, Bitstamp.
  • LATAM: dLocal, Bitso, Lemon Cash.
  • Africa: Yellow Card, Onafriq, Flutterwave.
  • Asia: varies widely by country.

For a B2B product targeting cross-border, you need at least one partner per major corridor you support. Onboarding a partner takes 2–6 months and meaningful legal/operational work. We typically recommend launching with three corridors and expanding from there.

The “did the payment land?” problem

The single most operationally painful part of running a stablecoin payment product is reconciliation. A user expects to receive funds, you initiated a transfer, but the funds haven’t shown up — what happened?

The possibilities multiply:

  • The on-chain transfer is pending (usually fine, just wait).
  • The on-chain transfer failed (insufficient gas, contract revert).
  • The off-ramp partner has the stablecoins but hasn’t initiated the fiat leg.
  • The off-ramp partner initiated the fiat leg, but the bank rejected it (wrong details, sanctions block).
  • The receiver’s bank credited the funds but hasn’t notified them.
  • Something in your reconciliation system is wrong.

The system needs to track every payment through every state across every partner, surface the current status to the user clearly, and have ops tooling to investigate stuck payments. This is more engineering work than the on-chain primitives themselves.

We recommend building a unified payment state machine from day one, with each partner integration emitting events into it. Don’t let each partner be a separate set of states the user has to understand.

What’s actually working in the market

A few patterns that are paying back well in 2026:

Cross-border payroll for distributed teams. Companies paying contractors and employees across borders. Traditional cost: 2–8% per payment plus delays. Stablecoin alternative: 0.5–1.5% with same-day settlement. Annual volume per company is large; pricing power is real.

Marketplace payouts. Platforms paying sellers across borders. Same dynamics as payroll, with the added benefit that the platform can hold the stablecoin in escrow during disputes.

Treasury management for crypto-native businesses. Companies with on-chain revenue paying off-chain expenses. Margin opportunity in providing the conversion + fiat banking layer.

B2B invoicing for international trade. Importers/exporters where the alternative is a 5–7 day wire with $50+ fees. Stablecoin rails compete strongly here.

What’s not working as well: consumer payments where the user just wants to pay for coffee. The friction of holding stablecoins is too high, the alternatives (Venmo, cards) are too good, and the regulatory overhead doesn’t pay back.

Where to start

If you’re building a B2B stablecoin payments product, the order we’d suggest:

  1. Pick one corridor and one product (e.g., US-to-Mexico marketplace payouts). Don’t try to be cross-border-payments-as-a-service from day one.
  2. Partner with a licensed entity for the regulated parts. Don’t try to be Circle.
  3. Build the unified payment state machine first. Everything else gets bolted onto it.
  4. Plan compliance staffing and tooling from week one. The ops cost is the biggest line item.
  5. Have legal counsel review your structure before you ship. The regulatory landscape shifts quarterly.

If you’re scoping a stablecoin payment product and want a second opinion on the architecture and partnerships, we’d be glad to dig in.