TL;DR

  • Your license path full bank charter, EMI/PI, or sponsor-bank BaaS decides your cost, timeline, and what you are legally allowed to offer.
  • A UK Authorised EMI needs £350,000 in capital; a full US bank charter has no fixed floor and can run from tens of millions to $210M+ depending on the business model.
  • BaaS is the fastest route (weeks to months) but gives you the least control; a full charter is the slowest (12–24+ months) but gives you full banking powers.
  • US regulators have opened up; the OCC received about 40 de novo charter applications in 18 months through mid-2026, up from historic lows, but the application quality bar has not dropped.
  • Compliance ownership stays with you no matter who builds your tech; outsourcing infrastructure never outsources regulatory accountability.
  • Most teams land on a hybrid: licensed core banking infrastructure plus custom product and UX, rather than pure build or pure BaaS.

If you are looking to figure out how to start a digital bank, the first thing to acknowledge is that this is not an app development decision. Rather, it’s a regulatory decision that comes by chance with an app. Most founders start by researching tech stacks and only later find that the license they choose decides almost everything else – what you can offer, how much capital you need, and how long it takes to make your business operational.

This guide is for founders, operators, and product leaders who already understand the fintech ecosystem and are beyond the “what is a neobank” stage. Whether you are deciding between a full bank charter, an EMI license, or just partnering with a bank through BaaS, either way, you want real numbers, not sales pitches, and you want to know what actually holds these launches up.

Quick Answer: How Do You Start a Digital Bank?

You start by picking a license path – full bank charter, a sponsor-bank/Banking-as-a-Service partnership, or an EMI/PI license because that pick sets your budget requirement, product ceiling, and timeline.

From there, you pick your target customers, set up a core banking system (built or licensed), add KYC/AML and card issuing, and launch your MVP. Your compliance team needs to be running by then, not still being hired.

Digital Bank vs. Neobank vs. BaaS Fintech: What’s the Difference?

Experts use these terms interchangeably; that’s why the planning stage is confusing.

Below are the practical differences:

Term What it actually means Who holds the license
Digital bank A deposit-taking institution that operates without physical branches, under its own or a partner charter Itself (full charter) or a partner bank
Neobank A brand and app layer offering banking-like features – usually the front end of a BaaS arrangement A sponsor bank behind the scenes
BaaS fintech A product built on top of a licensed bank’s infrastructure via APIs, no charter of its own The BaaS provider’s partner bank

If you are specifically preferring the neobank route, building the app and customer experience on top of someone else’s license – our guide to neobank app development covers the product and feature part of that decision in depth.

Digital Bank vs. Neobank vs. BaaS Fintech_-What's the Difference

Step 1: Choose Your License Path

Your license path is the only decision that determines the cost, timeline, and what you are legally allowed to do. Three practical routes in 2026:

License Path Typical Capital Timeline What You Can Do
Full bank charter (US de novo, national) Highly variable by business model – recent OCC approvals have ranged from tens of millions up to $210 million in initial paid-in capital for a full-service digital-asset-focused bank Often 12–24+ months, including pre-opening requirements Take deposits, lend from your own balance sheet, full banking powers
EMI or PI license (UK/EU) £350,000 initial capital for a UK Authorised EMI under the Electronic Money Regulations 2011; lower or no fixed minimum for Small EMI/PI if you stay under £5M e-money and £3M/month in transactions Roughly 6–12 months end-to-end Issue e-money, hold safeguarded funds, provide payment services – no own-balance-sheet lending or deposit-taking
Sponsor-bank / BaaS partnership No license capital of your own – you pay integration and revenue-share costs instead. Weeks to a few months to integrate Offer banking-like features under the sponsor’s charter, with less control over product terms.

If you are aiming at a US charter, the regulatory factor shifts in your favor. The Office of the Comptroller of the Currency (OCC) received about 40 startup bank applications over 18 months through mid-2026, compared to fewer than four submitted between 2011 and 2014. Today, it’s deciding various complete applications within about 120 days. That’s the real change in requirements from the regulator side. But this doesn’t mean the count of complete, credible applications has dropped.

On the EMI side, ensure that the UK’s protection rules have also changed: the Financial Conduct Authority’s (FCA) Policy Statement PS25/12 introduced a complete interim framework – the Supplementary Regime – effective from May 7, 2026, tightening how payment and e-money institutions safeguard customer funds.

If you are building for a UK EMI, your safeguarding architecture design should follow the latest rule, not the pre-2026 version.

Step 2: Define Your Target Market and Business Model

Go for a narrow segment before picking features. A digital bank targeting freelancers has a distinct KYC friction tolerance, card program requirements, and unit economics than one developed for SMB payroll or migrant remittances. The license you opted for in Step 1 also restricts this ahead – an EMI can’t originate loans from its own balance sheet, so a lending-led model generally requires either a banking charter or a lending partner layered on top.

Be genuine about the revenue early. Interchange alone rarely covers customer acquisition costs for a card-led consumer product. Most digital banks that survive two years add another revenue stream – subscriptions, SME services, embedded lending, or FX spread where they are licensed to offer it.

Step 3: What Does It Actually Cost to Start a Digital Bank in 2026?

Cost splits into two parts that founders generally blend and should not: regulatory capital and everything else – compliance and legal staffing, operations, and technology.

  • Regulatory Capital

As illustrated above, this ranges from efficiently zero (BaaS) to £350,000 (UK EMI) to tens or hundreds of millions of dollars for a full US charter, depending on your business model and risk profile.

  • Technology Build

A custom core banking mobile/web app, system, or ledger generally runs from the low hundreds of thousands into seven figures, depending on how much you build versus license. Licensing an existing core banking platform compresses this notably compared to building a ledger from scratch.

  • Operational Runway

Preopening capital needs for a US charter usually include enough runway to cover the bank’s first 1-2 years of operating losses before it needs to prove profitability.

  • Compliance and Legal

Budget for a compliance officer or team from day one, ongoing KYC/AML tooling, and legal counsel through the application process. This is not a one-time expense – it’s a permanent line item.

If you want a closer look at how such technology costs break down by app type – digital banking, P2P lending, or eWallet – our fintech app development guide goes into the feature-by-feature pricing.

How to Launch a Digital Bank

Step 4: Build the Core Banking and Technology Stack

Whichever license you choose, the technology layer looks extensively similar: a KYC/AML engine for onboarding and ongoing maintenance, a ledger that tracks each balance and transaction, a mobile or web front end customers actually use, and card issuing infrastructure if you are offered cards.

Here, the build-vs-buy decision matters more than most founders expect. A completely custom core banking system offers you lasting control, puts integration risk on your team, and adds months to your timeline. A licensed core banking platform gets you to MVP faster but partly ties your roadmap to a vendor’s release cycle. Most successful digital bank launches in 2026 land somewhere in the middle – licensed core infrastructure with custom product and UX on top.

Step 5: Who Owns Compliance? (You Do)

Outsourcing your technology stack doesn’t outsource your regulatory responsibility. This is the most common mistake founders make: they outsource infrastructure to a vendor and assume compliance comes bundled in. But it doesn’t. Your regulator holds you accountable for KYC/AML effectiveness, protection of customer funds, and fraud controls, regardless of who built the underlying system.

Recent OCC denial published as part of the agency’s new transparency initiative starting in 2026 aims at a consistent pattern: applicants who presumed they would finish compliance infrastructure after approval, despite having it substantially built before applying. Build your compliance function, real documented procedures, real staff, in parallel with your technology, not later.

Step 6: Launch Your MVP and Track the Metrics That Matter

Once you are live, the metrics that tell you whether your business works are distinct from the vanity metrics most product teams default to. Analyze:

  • KYC Pass Rate: How many applicants clear onboarding without manual review
  • Activation Rate: The share of signups who fund an account and transact
  • Payment Success Rate: Failed transactions erode trust rapidly in a banking product
  • Reconciliation Breaks: Mismatches between your partner bank’s records and your ledger, which point to integration issues before customers notice them.
  • Fraud Loss Ratio: A leading indicator of whether your risk controls are working effectively.
  • Compliance Rate: Often the earliest signal of UX or policy issues, well before it impacts churn figures.

How to Launch a Digital Bank

Build In-House, Partner, or Buy? Where Nimble AppGenie Fits

None of these three paths – full custom build, pure BaaS integration, or licensed core banking platform – is universally right. A well-funded team chasing a full charter generally needs custom infrastructure they control end to end. A leaner team validating a niche segment is usually better served starting on BaaS and migrating to more control later, once the model is proven.

Nimble AppGenie works across all three models: custom core banking builds for teams pursuing their own charter, neobank app development for BaaS-based launches, and integration work for teams migrating between platforms as they scale. If you want a second opinion on which model fits your specific business plan for banking software development services, book a call with our fintech team before you lock in a technology direction.

 

FAQs

It depends entirely on your license path. A UK Authorised EMI requires £350,000 in initial capital. A full US bank charter has no fixed floor, but recently OCC approvals for full-service banks have required initial paid-in capital ranging from tens of millions of dollars up to $210 million, depending on the business model and risk profile. A BaaS partnership requires no license capital at all, though you’ll still need runway for technology and operations.

Yes. Most digital banks in the market today operate under an EMI/PI license or a sponsor-bank BaaS arrangement rather than a full charter. Both let you offer deposit-like accounts and payment services, but neither lets you lend from your own balance sheet or take true deposits the way a chartered bank can.

A BaaS-based launch can go live in a few months once integration is complete. A UK EMI application typically takes 6 to 12 months end to end. A full US bank charter is the longest path, commonly 12 to 24 months or more, including pre-opening requirements, even though the OCC has been deciding a growing share of complete applications within about 120 days in 2026.

An EMI can issue electronic money, hold customer funds in safeguarded accounts, and provide payment services, but it cannot take deposits in the legal sense or lend from its own balance sheet. A full banking charter allows both, at the cost of significantly higher capital requirements and a longer, more demanding application process.

Regulatory appetite has clearly increased: the OCC received around 40 de novo applications in the 18 months through mid-2026, compared with historic lows earlier in the decade, and has been approving a meaningful share of them. That said, the OCC has also signaled it’s scrutinizing application quality more closely, with published denials pointing to unfinished compliance and operational work as the main reason applications fail. Increased appetite doesn’t mean a lower bar for readiness.