Briefing · 9 min read

Most fintech licences fail at the bank, not at the regulator

Most fintech founders assume the regulator is the hard part. In practice, the FCA, MFSA, DFSA or CBB will approve on paper — and the acquiring bank will then decline on evidence. Safeguarding accounts, settlement banking and card-scheme sponsorship are where EMI and PSP applications quietly collapse. This is what a bankable licence file actually looks like in 2026.

Meridian Editorial28 June 2026210 views
Most fintech licences fail at the bank, not at the regulator

A licensed fintech that cannot open a bank account has a business problem, not a licensing one — and it happens often enough that the pattern is now familiar. This briefing sets out the structural reasons EMI, PI and small-bank licences run into trouble with correspondent banking, and how to design a licence application with bankability in mind from day one.

The landscape in 2026

The EU's EMI/PI framework, the UK FCA regime, the UAE (CBUAE and VARA), Singapore (MAS PS Act) and various offshore centres all license non-bank payment institutions. Post-Wirecard, the tier-one correspondent banks (BNY, Citi, JPM, Deutsche, HSBC) have narrowed the population of EMIs they will bank, and mid-tier correspondents have followed. A licence in a jurisdiction the correspondents don't like is a licence you can't operate.

Why banks refuse

The refusals cluster around a few themes:

  • Jurisdiction of licence. Some jurisdictions have too many licensees, too little supervision, or a track record of failures. Lithuania, Malta and Cyprus went through this in different periods. Even a good operator based in the wrong jurisdiction pays a premium.
  • Business model risk. Cross-border, multi-currency, crypto-adjacent, gaming-adjacent and high-risk-merchant portfolios all reduce correspondent appetite.
  • Ownership and governance. Ultimate owners in high-risk jurisdictions, PEP exposure, and boards without banking experience all show up in correspondent onboarding.
  • Compliance stack. Off-the-shelf transaction monitoring, thin sanctions screening, no local MLRO — correspondents look for real capability, not a checkbox.
  • Capital adequacy. A licence issued at minimum capital may pass the regulator but fails the bank's tier-one counterparty policy.

Designing for bankability

The licence application should be built with the future bank onboarding pack in mind. In practice that means:

  • Pick the licensing jurisdiction based on where the correspondents want to see you — not where the licence is cheapest. Ireland, Luxembourg, Germany and (for UK-scope) the FCA are the strongest current picks for EMI/PI.
  • Capitalise above the regulatory minimum. Tier-one correspondents look for real capital, not just the €350k EMI floor.
  • Recruit a board and MLRO with banking-grade CVs. This is the single most-cited factor by correspondents.
  • Choose a tier-one transaction monitoring vendor and be able to demonstrate model validation and tuning.
  • Segment your business plan: define the corridors, currencies and merchant categories, and evidence the AML controls appropriate to each.

Route to a working correspondent stack

Realistically, most EMIs and PIs cannot open a tier-one correspondent account on day one. The pragmatic route is:

  1. Sponsor bank or BaaS provider for launch. A licensed sponsor lets you go live while the direct correspondent relationships mature.
  2. Second-tier correspondent for volume. Typically a specialist Baltic, German, Dutch or UK bank willing to serve licensed EMIs with real controls.
  3. Tier-one correspondent for scale. Usually 18–24 months into operation, with a clean audit history and demonstrable AML performance.

Trying to skip steps is the most common cause of expensive licence pauses.

How we approach this at Sovereign Signal

We work with fintech founders at the licence design stage, before the application is filed, to make bankability part of the plan. We introduce sponsor banks and correspondent candidates we have worked with directly, help prepare the KYC file the banks will actually read, and coordinate with the regulatory counsel running the licence process. We don't take referral fees from banks; that would compromise the recommendation.

Worked example

An EMI applicant proposed to launch in Malta with the minimum capital and a crypto-heavy client list. The Maltese licence was achievable but no tier-one or tier-two correspondent would bank it. We reworked the plan: Irish CBI licence, €2m capital, non-crypto merchant scope for year one, a tier-two Dutch correspondent lined up in advance, and a UK BaaS provider for immediate launch. The licence took six months longer; the business is now operational and profitable with two direct correspondents.

FAQs

Which jurisdictions are correspondent-friendly for EMIs in 2026?

Ireland, Germany, Netherlands, Luxembourg and — for UK sterling — the FCA regime. Lithuania and Malta remain licensable but bring correspondent friction.

Can I use a BaaS provider indefinitely?

Yes, but at a cost — margins are thinner and you carry the provider's own risk. Most scaling EMIs move to direct correspondents within two years.

How much capital should I actually put in?

For a serious EMI with correspondent ambitions, €1–3m is realistic, well above the €350k regulatory floor.

Do I need a physical office in the licensing jurisdiction?

Yes. Every major regulator now expects real presence, and correspondents check.

Related reading

Stay ahead of regulation changes

Get an email when the rules move — only on the topics you choose.

Pick what you care about. We'll send a short note when a jurisdiction, treaty or regulator update actually changes what you should do.

Topics

One-tap unsubscribe on every email. We never share your address.