Embedded finance solutions are the APIs, sponsor-bank relationships, card-network connections, and settlement rails that let a non-financial software product offer accounts, payments, cards, credit, or cross-chain money movement to its users under its own brand. Bain and Company and Bain Capital project revenue for platforms and infrastructure providers will more than double from $21 billion in 2021 to $51 billion in 2026, on roughly $7 trillion in US transaction value (Bain). This guide covers the four solution categories, a worked example of one program from scratch, a decision framework, and where stablecoin rails fit next to cards and ACH.
What are embedded finance solutions?
Embedded finance solutions are the packaged infrastructure a software company buys instead of building. The vendor holds the regulatory license, the bank contract, the card-network membership, or the onchain settlement logic. The software company integrates through an API and ships the resulting account, card, payment, or transfer to its own users under its own brand. That is what turns an expense product into a card program and a marketplace into a payout engine.
The category exists because the alternative is slow. Getting sponsored by a bank, integrating with card networks, standing up KYC, monitoring transactions for fraud, and reconciling ledgers in-house takes years and a compliance team. Ramp launched its first corporate card program in under 70 days working with Marqeta as the issuing platform (Marqeta case study). Programs like that are the reason embedded finance became a distinct software category rather than a feature bolted onto banking.
The four categories of embedded finance solutions
Every embedded finance solution slots into one of four buckets: banking-as-a-service, payments-as-a-service, issuing-as-a-service, and stablecoin-rail-as-a-service. Most real programs stitch two or three together because the underlying rails do different jobs. Banking-as-a-service moves fiat between US bank accounts. Payments-as-a-service accepts cards and pays out sellers. Issuing-as-a-service creates cards on the platform's brand. Stablecoin rails move value across blockchains without a bank in the middle.
Banking-as-a-service (BaaS)
BaaS wires a sponsor bank behind an API so the platform can open FDIC-insured accounts, hold customer balances in a for-benefit-of structure, and move money over ACH, wire, RTP, and FedNow. Unit is the archetype for US software platforms and reports over $100 billion in annualized transaction volume, more than 2 million users, and 11 million daily API calls on its homepage (Unit). Unit itself is not a bank; banking services come from its FDIC-member partners. Synctera plays a similar role but pairs fintechs with sponsor banks in a shared oversight console so both sides see the same compliance data (Synctera). Treasury Prime differs by keeping the fintech in direct contract with a partner bank rather than acting as program manager (Treasury Prime).
Payments-as-a-service (PaaS)
PaaS accepts money from a customer and pays out to a seller, driver, contractor, or creator, and it handles onboarding, KYC, sanctions checks, tokenization, and dispute workflows so the platform does not. Stripe Connect is the default at platform scale and is what Shopify, DoorDash, Lyft, Instacart, and GitHub use to route funds to their sellers, drivers, and creators (Stripe Connect). Adyen offers the same shape but on its own licensed banking infrastructure, so a global platform can white-label payments, accounts, card issuing, and capital without stitching together separate contracts (Adyen). Rapyd leans into cross-border acceptance, multi-currency accounts, and local payout methods, with Uber, Adidas, Microsoft, Booking, Google, and Meta named on its homepage (Rapyd).
Issuing-as-a-service (IaaS)
IaaS lets a platform create physical and virtual cards on its brand with authorization logic the platform controls. Marqeta is the reference platform, with an open API for card issuing and real-time processing, Just-in-Time funding, and dynamic spend controls that let a program approve or decline authorizations based on live business logic (Marqeta). Highnote combines issuing, acquiring, credit, and real-time ledgering under a single GraphQL API with money-movement rails spanning ACH, wire, RTP and FedNow, OCT and AFT, and stablecoin (Highnote). SoFi Tech Solutions, the former Galileo, powers cards and digital banking for larger programs and rebranded under SoFi in 2025 (SoFi).
Stablecoin-rail-as-a-service
Stablecoin rails move dollar-denominated value across blockchains without touching card networks or ACH. Circle's Cross-Chain Transfer Protocol enables native USDC to move 1:1 across supported chains through a burn-and-mint mechanism, with Fast Transfer for sub-finality settlement and Hooks for automated post-transfer actions (Circle CCTP). CCTP supports Arbitrum, Avalanche, Base, Ethereum, Solana, Polygon, OP Mainnet, Unichain, and roughly two dozen other chains. Stripe put stablecoins on the same API surface as cards by completing its acquisition of Bridge in February 2025, folding stablecoin infrastructure into Stripe's platform for global money movement (Stripe). Eco Routes sits one layer up: an intent-based product where a user signs the outcome they want, solvers compete to fulfill it, and a source-chain vault releases the reward only on cryptographic proof of fulfillment (Eco docs).
A worked example: how a spend-management product ships a card
The clearest way to see how embedded finance solutions fit is to walk through a real program end to end. Ramp is a spend-management SaaS that offers corporate cards, expense automation, bill pay, and travel to businesses. The card program runs on Marqeta as the issuing platform. According to Marqeta's case study, Ramp went from incorporation to its first transaction in under 70 days on Marqeta's infrastructure and can spin up single-use virtual cards, physical cards, and unlimited card counts per customer with detailed rules around where, when, and how often each card can be used (Marqeta).
What Ramp did not build: the card BIN, the network connection to Visa or Mastercard, the authorization stream, the tokenization service that powers Apple Pay and Google Pay, the PCI-compliant card-data storage, or the bank partnership that funds the settlement account. What Ramp did build: the customer onboarding, the spend-policy engine, the receipt-matching workflow, the accounting integrations with NetSuite and QuickBooks, and the UI for admins and cardholders. Marqeta's Just-in-Time funding lets Ramp approve or decline each transaction in the authorization stream based on Ramp's own policy engine, which is the mechanical reason Ramp can enforce spend policies at the moment of purchase rather than after the fact (Marqeta platform).
That is the pattern. The platform buys the parts that require regulatory infrastructure, network membership, or bank contracts, and it builds the parts that are its actual product: policy, UI, workflow, and integrations. The vendor selection is not "which brand" but "which layer, which rail, and which contract model."
Which questions should a team ask when picking an embedded finance solution?
The right vendor depends on five concrete decisions, not on brand recognition. Each decision maps to a technical or contractual dimension the platform will live with for years. Answer these first, then pick the vendor whose model matches. Ranking vendors before answering these usually produces a bad fit that the team discovers six months into integration.
1. Which rail does the money need to cross? Card networks for cardholder-present or online purchases. ACH and wire for US bank-to-bank money movement. RTP and FedNow for instant fiat transfers under a per-transaction cap. Local rails for cross-border acceptance in countries where cards are not the default. Stablecoins for value that moves across chains or across borders where local banking rails are slow or expensive. The rail decides the vendor shortlist, not the other way around.
2. Who holds the bank contract? In a program-manager model the tech vendor sits between the fintech and the bank, and the fintech contracts with the tech vendor. In a bank-direct model like Treasury Prime the fintech signs and manages the bank relationship itself while the tech vendor provides the software layer (Treasury Prime). Bank-direct usually gives the fintech better deposit economics and more room to negotiate; program-manager is faster to launch. Regulatory scrutiny on BaaS structures increased through 2024 and 2025, which is part of why Synctera's shared-oversight console and Treasury Prime's bank-direct model gained traction versus pure program-manager setups.
3. Who owns compliance responsibility? Some vendors run KYC, sanctions screening, and transaction monitoring in-house and expose the results through the API. Others require the platform to plug in its own KYC and monitoring vendors. This decides how big a compliance team the platform needs on day one and how much of the ongoing operating cost lives with the platform versus the vendor.
4. What latency does the product need? Card authorizations settle in milliseconds. ACH clears in one to three business days. Wire clears same-day. RTP and FedNow settle in seconds under a cap. Circle's CCTP Fast Transfer settles USDC across chains faster than finality on the source chain (Circle CCTP). Match the rail to the product promise, not to the rail the team already knows.
5. What is the fee structure and where does the platform make money? Card interchange, ACH per-transaction fees, monthly platform fees, share of float on customer deposits, and stablecoin gas fees each have different economics. A platform whose value proposition is float on deposits will make different vendor choices than one whose value proposition is interchange on card volume.
When stablecoin rails fit next to traditional rails
Stablecoin rails do not replace cards or ACH; they cover the flows those rails cover badly. Card and ACH rails are the right answer when the counterparty is a US bank account or a Visa or Mastercard credential and the value stays in dollars in the domestic banking system. Stablecoin rails become the right answer when value moves between chains, between exchanges, or across borders in a currency the local banking system does not clear cheaply. That is the underlying reason Stripe added stablecoin infrastructure to its platform after acquiring Bridge, and why Highnote lists stablecoin alongside ACH, wire, RTP, and FedNow as a money-movement rail (Highnote).
The concrete pattern for a platform that already has a card or ACH program: keep those rails at the customer edge and use a stablecoin rail underneath for treasury, cross-border settlement, and reconciliation across geographies. Eco Routes handles the cross-chain side of that stack. A user or program signs the outcome it wants, solvers compete to fulfill it, and a source-chain vault releases the reward only on cryptographic proof of fulfillment across CCTP, Hyperlane, LayerZero, Chainlink CCIP, or Polymer, with the ERC-7683 cross-chain order interface as the standard contract shape (Eco docs). The platform does not have to pick a single bridge; it commits to an outcome and the solver network picks the rail per transfer.
What does implementing an embedded finance solution actually look like?
Implementation splits into four surfaces the platform's engineering team touches. The API surface, the identity handoff, the ledger reconciliation loop, and the operational tooling. Each of these has a known shape regardless of which vendor the team picks, and estimating them accurately before signing is what separates a 90-day launch from an 18-month one. The specifics below map to a card-program build; a BaaS or stablecoin build has the same four surfaces with different endpoints.
API surface. A modern issuing or BaaS vendor exposes REST or GraphQL endpoints for creating customers, opening accounts, issuing cards, authorizing transactions, moving money, and fetching statements. Marqeta's Core API covers cardholders, cards, funding sources, transactions, and webhooks; Highnote unifies issuing, acquiring, credit, and ledger under one GraphQL endpoint (Highnote). The platform decides whether to call these endpoints server-side from its own backend or embed vendor-hosted UI components for onboarding, KYC, and card display. Vendor-hosted components reduce PCI scope and shorten the launch timeline.
Identity and KYC handoff. Every regulated flow requires the platform to prove the end user is who they say they are. In a program-manager model, the vendor's KYC vendor runs the checks and returns pass, fail, or manual-review. The platform stores the resulting identifier and passes it back on every subsequent API call. Failures fall into a manual-review queue that a compliance operator works through in the vendor's console. This is where launch timelines slip when the platform underestimates the manual-review volume.
Ledger reconciliation. The platform holds its own ledger of what it thinks happened, and the vendor holds its ledger of what actually settled. Once a day, the two are reconciled through a webhook stream or a batch report and the differences are booked. This is the least glamorous surface and the one that catches production bugs. A platform that skips the reconciliation build ends up with unexplained balance differences after ninety days and no way to answer a customer support ticket about a missing payment.
Operational tooling. Support agents need to look up transactions, freeze cards, and refund charges. Compliance operators need to work KYC exceptions and dispute cases. Finance needs to close the books. Vendor consoles cover most of this, but any nontrivial program builds an internal console on top of the vendor API so agents do not have to switch between the vendor's console and the platform's own product for every ticket.
Vendor landscape reference
A full vendor-by-vendor breakdown, including who is out of the market in 2026 (Bond, folded into FIS; Solid, Chapter 11), lives in a companion piece. See Embedded Finance Companies 2026: The Landscape for the ten vendors currently worth evaluating and how each sits on the four-category grid above. The categories in this guide map one-to-one onto the vendor list in that piece.
FAQ
What is the difference between banking-as-a-service and payments-as-a-service?
Banking-as-a-service exposes a sponsor bank's balance sheet through an API: accounts, ACH, wires, sometimes savings. Payments-as-a-service accepts money from a customer and pays out to a seller, driver, or creator, and it handles the acceptance rails, dispute flow, and payout rails end-to-end. Some vendors like Adyen and Unit cover both under one contract.
Do I need a sponsor bank if I use an embedded finance solution?
Yes, for any US-regulated flow. The BaaS vendor either supplies the sponsor bank as part of the package (Unit, Synctera) or the platform contracts with the bank directly while using the vendor's software layer (Treasury Prime). Card issuers like Marqeta and Highnote also run on partner bank BINs. There is no US embedded finance program that does not sit on a bank somewhere.
How does a stablecoin embedded finance solution compare to a card program?
A card program uses interchange, Visa or Mastercard, and a partner bank BIN, and settles into the platform's operating account in one to two business days. A stablecoin program uses onchain settlement, no card network, and settles across chains in seconds to minutes depending on the rail. Card programs fit customer-facing spend; stablecoin programs fit cross-chain treasury and cross-border settlement.
What does an embedded finance integration cost?
Vendor pricing is usually a mix of platform fees, per-transaction fees, and revenue share on interchange or float. Public rate cards are rare; most vendors quote based on projected volume. Engineering cost on the platform side runs from three to nine months for a first card or account program, driven mostly by the identity handoff and reconciliation build.
Is Bond still a viable embedded finance solution?
No. Bond was acquired by FIS in June 2023 and folded into FIS's broader offering, so it is no longer an independent vendor. Solid Financial Technologies filed Chapter 11 in April 2025. Older roundups still list both, which is why 2026 evaluations should be cross-checked against the current landscape piece linked above.
Related reading
Best Treasury Management Software 2026, the 10-vendor listicle comparing enterprise, mid-market, and operating-cash treasury platforms with real pricing bands.
Sources and methodology. Vendor descriptions in this piece were verified against each company's own website, docs, or a primary case study on 2026-08-13. Market sizing comes from Bain and Company and Bain Capital's 2022 embedded finance report. The Ramp and Marqeta implementation timeline is from Marqeta's published Ramp case study PDF. Stablecoin rail claims are drawn from Circle's CCTP product page and Stripe's Bridge acquisition newsroom post. Figures on this page refresh when the underlying primary sources publish updated numbers.
International payment gateway explained walks through vendor pricing and a worked example.

