Every team that moves money eventually hits the same fork. Do we get authorized ourselves, sit under someone else's authorization, or hand the regulated parts to a partner and build a product on top. The mistake we see most often is treating this as one decision. It is three, and they come apart.
This module ties together the earlier ones. The license question, renting a license, safeguarding, treasury, cross-border, and reconciliation all feed a single operating choice: where does the regulated surface sit, and how much of it do you own. Get the framework right and the rest of the stack falls into place.
Three decisions, not one
"Build vs rent vs partner" is shorthand for three separate questions that people collapse together.
The first is licensing. Do you hold the authorization that lets you provide payment services or issue e-money, or do you operate under someone else's. The second is accounts. Who holds client funds, where are they safeguarded, and whose name is on the account at the bank. The third is processing. Who runs the rails, the BIN, the message formatting, and the settlement.
You can mix and match. A common pattern is to rent the license, partner for processing, and build the product and ledger yourself. Another is to hold your own license but partner for card issuing through a BIN sponsor. Treating the three as one bundle is how teams either over-build, paying for an authorization they did not need, or under-build, discovering in year two that their economics are owned by a partner.
What each path actually costs
Owning a license is the heaviest path and the one people underestimate. For a UK Authorised Electronic Money Institution, the FCA application fee is £5,580, initial capital is £350,000, and authorization typically takes 6 to 18 months, longer than a payment institution because the FCA scrutinizes safeguarding methodology and float projections. Once live, ongoing own funds run at the higher of the £350,000 floor or 2 percent of average outstanding e-money, a number that grows with your float.
Total first-year spend, including legal, compliance hires, and technology, commonly lands between roughly £750,000 and £1.2 million. That is before you have processed a single transaction. The reward is control: your economics, your client relationships, your roadmap, no principal who can change your terms.
Renting a license, operating as an agent or distributor under a principal EMI or PI, removes the capital and timeline problem. You can go live in weeks rather than quarters. The trade is structural. You act under the principal's authorization and their responsibility, which means their risk appetite caps your product, and their pricing sits inside your margin. We covered the mechanics in the renting module; the point here is that renting is a speed-and-capital decision, not a permanent one.
Partnering for processing and accounts is the third lever. A BIN sponsor, a regulated principal member of Visa or Mastercard, lets you issue cards without scheme membership of your own. A banking partner holds and safeguards client funds. You build the product surface and own the customer, but the regulated plumbing belongs to someone else.
A worked example
Take a B2B expense card startup in the UK that wants to launch in one quarter.
Building the full stack means an AEMI license at £350,000 initial capital, a 6-to-18-month wait, plus scheme membership. The math kills the timeline before it starts. So they decompose. They rent the license by operating as an agent of a principal EMI, which puts a regulated entity on the hook for safeguarding and lets them launch in weeks. They partner for processing through a BIN sponsor so cards work on day one. They build the ledger, the approval logic, the reconciliation, and the dashboards, because that is the product and the moat.
Eighteen months in, the picture changes. Float is large enough that 2 percent ongoing own funds is cheaper than the principal's per-transaction take, and the principal's risk policy is blocking a product they want to ship. Now the license math flips, and they apply for their own authorization while keeping the BIN sponsor. The accounts and processing decisions did not have to move just because the licensing one did. That is the framework working as designed.
The questions that decide it
Four questions usually settle the call.
How fast do you need to be live?
If the answer is weeks, you are renting and partnering, full stop. Authorization timelines do not bend for a launch date.
Where does float economics break even?
Run the number. Once 2 percent ongoing own funds plus license overhead is cheaper than the per-transaction or revenue share a principal takes, owning the license starts paying for itself. Below that line, renting is simply cheaper.
Who controls the parts of the product you cannot compromise?
If a partner's risk appetite or roadmap blocks something core to your business, that regulated surface needs to move in-house. If it does not touch the thing customers pay for, leave it with the partner.
Can you carry the compliance load?
Owning a license means owning safeguarding, AML, reporting, and wind-down planning. The UK safeguarding regime is tightening under the FCA's PS25/12, with strengthened rules taking effect 7 May 2026. If you cannot staff that properly, a principal carrying it for you is worth paying for.
The closing takeaway
Decompose before you decide. Map licensing, accounts, and processing as three independent dials, then set each one against speed, float economics, control, and compliance capacity. Most teams should rent and partner early, build the product and ledger that differentiate them, and move regulated surfaces in-house only when the float math, the control math, or both clearly justify it. The right answer is rarely all-build or all-partner. Treat it as a mix you revisit as you scale.