Most fintech founders are quoted a licensing project before anyone has asked whether they need one. The first question is not how to get licensed. It is whether you are transmitting money, lending, or neither.
The usual trigger. If funds arrive from one person and you are responsible for getting them to another, you are in the analysis whether or not you call it a transfer.
A balance a customer can draw down later is a liability you owe them. Most states treat that as money transmission regardless of how the product is marketed.
Control matters more than custody. If you can decide where money goes while it is in motion, that is usually enough.
Passing instructions to a bank or processor that moves the money itself may fall outside the definition. May. It depends on the contracts and the account structure, not on the architecture diagram.
Most states exempt a business collecting on behalf of a merchant it genuinely represents. The exemption is narrower than founders assume and turns on whether the payer's obligation is discharged when you receive the funds.
The analysis turns on specifics most founders have never been asked about: whose name is on the account, who bears the loss if funds go missing, who the customer’s contract is with, when title to the money passes, and whether you can direct funds while you are holding them. A clear answer here is worth more than anything else on this page. It decides whether you need a multi-state licensing program, a sponsor bank, a narrower product, or nothing at all.
Registering with FinCEN as a money services business is a form you submit. It is not permission to operate, and nobody reviews your business before it takes effect.
Each state regulator approves you separately, on its own timeline, against its own requirements. Founders routinely complete the federal registration, see it accepted, and believe they are authorised. They are not.
Most states now scale it against total assets, so the requirement grows with the business. Model your capital requirement at the volume you are forecasting, not the volume you have.
Bond amounts rise with transmission volume and with the number of states. Some states accept a deposit or letter of credit instead.
States require licensees to hold eligible assets against outstanding obligations. This limits treasury strategy in ways that surprise finance teams late.
Capital requirements belong in the model before you raise, not after a regulator asks for an audited balance sheet.
The Conference of State Bank Supervisors published the Money Transmission Modernization Act to harmonise definitions, exemptions and net worth across states, and the large majority of money transmission activity now happens in states that have adopted some form of it.
States enacted different versions at different times with their own amendments, and interpretation already varies enough that CSBS opened a formal guidance-request process to deal with the divergence. A handful of states remain outside the model entirely.
Licensing every state at once is rarely the right first move. Where your customers actually are, which regulators move fastest, and which have requirements you cannot meet yet — that ordering is the difference between launching in a year and launching in three.
It is normal for a program to launch without a handful of states and add them later. It is better to plan that deliberately than to discover it during onboarding.
State lender licensing is the direct route. The bank partnership model is the common alternative, and it is the one that carries structural risk rather than administrative burden.
A California court ruled in 2026 that the partner bank rather than the fintech was the true lender in a well-known program. That decision is on appeal and binds nobody, and several states have true-lender statutes that reach further than the theory it rejected.
Several states have opted out of federal rate exportation for state banks, which is why most programs carry a state-by-state matrix of different terms, different products, or no availability at all.
Who funds the loan and from whose account, who holds it at origination, who keeps economic interest in the receivable, who sets credit policy and who can override it. A program where the bank’s role is real survives scrutiny. One where it is nominal does not, however the agreement is drafted.
Whether your product is money transmission, lending, both or neither — and what would have to change for the answer to change.
Agent-of-payee and payment processor exemptions, sponsor bank arrangements, narrowing the product at launch. Not every route is open to every business, and some that are marketed heavily do not hold up.
Which states, in what order, against where your customers are and what you can satisfy today.
Sponsor bank and program agreements, payment processor and banking-as-a-service contracts, merchant and platform terms, customer agreements, and the flow-of-funds provisions that set your regulatory position as much as the licensing analysis does. See vendor and customer contracts and SaaS and licensing.
A written anti-money-laundering program with a designated officer, customer identification, transaction monitoring, reporting and independent testing — the first thing both a regulator and a sponsor bank ask to see. See corporate compliance.
Sponsor banks, payment partners and investors all run regulatory diligence. Most early fintechs fail it on documentation rather than on substance. See strategic counsel.
Preparing and prosecuting dozens of NMLS applications is specialist work done by people who do nothing else. When a program reaches that point I bring those lawyers in and manage them alongside you, which is the same approach taken with IP, tax and immigration counsel.
A written program is not a compliance function. At volume you need someone running it day to day; the program and the policies are what they run.
The analysis that tells you whether you need the program, the structure that sets its scope, the contracts that create your regulatory position, the compliance infrastructure, and the ongoing general counsel role as the company grows into it. That is the work that is usually missing.
The expensive mistakes in this sector are made at the design stage, before anyone consults a lawyer. A flow of funds that creates an obligation nobody intended is cheap to fix in a design conversation and expensive to fix after launch.
Program agreements arrive on the bank’s form and set your economics, your compliance obligations and your exit terms at once.
If the product description you are raising on does not match what the system actually does with money, that gap is a regulatory problem and a diligence problem at the same time.
If you are pre-product or pre-revenue, the right engagement is a scoped analysis and a structure, not a fifty-state plan.
Digital asset businesses face an overlapping but distinct regulatory posture. Ask before assuming this page covers your situation.
Companies that clear the threshold analysis and start building usually move to an ongoing arrangement rather than returning per question. That is the Fractional General Counsel model, and in a regulated business it tends to pay for itself earlier than it does elsewhere.