4 Myths About Launching a Payment Business

Founder Perspective

What the market tells you. What actually happens. And why the gap between them costs founders anywhere from six months to a million dollars.

Key Realities at a Glance

• $600k–$1m: Real cost of buying an Electronic Money Institution (EMI), not the advertised $300k.
• 12–18 Months: Timeline until your first transaction when building from scratch.
• 6+ Months: Required for banking revalidation after an ownership change.

Every week I talk to founders who want to launch a payment business. Smart people. They've done research. They have a plan.

And almost every time, the plan is built on one of four assumptions that sound reasonable but aren't. Not because the information isn't out there — but because the people selling these shortcuts have no incentive to correct them.

So here's what the market won't tell you.

1. Myth 01 of 04

• What people think: "Buy an existing EMI for $300k and you're done."
• What's actually true: You're buying the right to start a conversation — not a ready-made infrastructure.

The $300k figure exists. You can find EMI entities at that price point — typically in jurisdictions with looser oversight, often with histories you can't fully verify. The logic seems obvious: existing licence, skip 12–18 months of build, pay and go.

Here's what that number doesn't include:

• Acquisition costs, legal due diligence, regulatory notifications
• Required capital buffers — the regulator doesn't care what you paid for the shell
• Full revalidation of every banking relationship from scratch
• FCA fit & proper assessment of you personally as the new controller

By the time you've done this properly, you're at $600k–$1m before a single transaction. And that's assuming the asset is clean — no off-balance sheet arrangements, no legacy compliance issues, no skeletons the previous owners left behind that the regulator hasn't spotted yet.

A reputable EU EMI — Netherlands, for example — starts at €1m for the entity alone, plus capital on top. The cheap options are cheap for a reason.

More importantly: the FCA isn't evaluating the company you bought. They're evaluating you. Your background. Your fitness and propriety as the new controller. A clean shell doesn't transfer credibility.

2. Myth 02 of 04

• What people think: "Buy one with existing banking accounts — they transfer with the entity."
• What's actually true: Banking relationships belong to the previous owners. They don't transfer. Ever.

This is a specific misunderstanding that costs people real money. The logic seems sound: the EMI already has correspondent banking relationships. Buy the EMI, inherit the accounts. Simple.

That's not how correspondent banking works.

The moment ownership changes, every correspondent bank reviews the relationship from zero. Not as a formality — as a genuine reassessment. New controllers. New business model. New risk profile. The accounts existed because of the previous owners: their transaction history, their volume, their risk appetite, their relationships with the bank.

You are not them.

The bank isn't transferring a relationship. They're deciding whether to start a new one — with you. That process takes 6+ months. And sometimes, after everything, the answer is no. Their risk appetite has changed. Your model doesn't fit the jurisdiction they want to focus on. The sector you operate in is currently under review.

• What you bought: An entity, a licence number, some filing history
• What you didn't buy: The correspondent bank's willingness to work with you
• Timeline until you know: 6+ months after close

It's not nothing. But it's not what most people think it is. And understanding the difference before you wire the money — that matters.

3. Myth 03 of 04

• What people think: "$100k lawyers, $500k capital, vibe-code the rest."
• What's actually true: You can't vibe-code regulatory infrastructure. The expert team costs the same whether you start right or start over.

I genuinely hope this works for someone. But here's what I keep thinking about.

You can't vibe-code banking keys. Or scheme keys. Or an HSM. These aren't features — they're cryptographic infrastructure sitting between you and every transaction your users will ever make. One wrong implementation and you're not just down. You're a fraud vector.

You can't vibe-code compliance logic that stays compliant. Regulations change. Edge cases compound. The logic that passes testing today fails an audit 18 months later — because no one truly understood why it was written that way. Because no one wrote it. Something generated it, and then moved on.

You can't vibe-code accountability. When a transaction fails at 2am and a client is screaming, who owns it? The model that generated the code is on version 4. The developer who "reviewed" it has moved on. You're left debugging something nobody truly understands — in a regulated environment where "we used AI" is not an answer the FCA accepts.

• Reliability in payments is not a nice-to-have. It's the product.
• Compliance logic requires expert maintenance indefinitely — not a one-time build.
• Regulatory accountability requires humans who own decisions, not models that generated them.

The teams that get this right spend just as much on expert humans as they would have spent doing it properly from the start. The ones that don't tend to discover this at the worst possible moment.

4. Myth 04 of 04

• What people think: "Just get our own licence from scratch — full control, cleaner long-term."
• What's actually true: £1.2m–£2m. Year one alone. 12–18 months before your first transaction. Your entire team in licence mode, not product mode.

This is the most defensible myth. Full control is real. Long-term ownership of your regulatory stack is genuinely valuable. For the right business, at the right stage, building from scratch is the right call.

But most founders who say this haven't done the full math.

• £1.2m–£2m: Year one cost for a UK EMI from scratch
• 100%: Percentage of your leadership team stuck in licence mode

The cost isn't just the legal and regulatory spend. It's the opportunity cost of 12–18 months where your entire leadership team is focused on getting a licence — not building the product, not acquiring customers, not learning what the market actually wants from you.

Competitors who launched on existing infrastructure are 18 months ahead. They've already found their first hundred customers. They know which assumptions were wrong. They've iterated.

You're still waiting for your first transaction.

• Licence-building is a full-time job for your entire leadership team
• Customer acquisition clock starts only after approval — not before
• 18 months is optimistic. Timeline slippage is common, not exceptional.

This path makes sense eventually, for the right business. It rarely makes sense at the start — before you've validated the model, before you know what you're actually building, before you have the revenue to justify the overhead.

The Bottom Line

The real question isn't which path is cheapest.
None of these paths are impossible. Some businesses have done all of them successfully. But the ones that succeed go in with accurate information — not the version that the people selling shortcuts want you to believe.

The question is which path gets you to your first transaction with the least wasted capital, the fewest regulatory surprises, and a team that's been building product rather than managing legal process. That calculation looks different for every business. But it looks better when you start with the real numbers.

Vladimir Dereviagin Co-Founder, Gemba