How Much It Costs to Launch a Neobank in 2026

When a fintech founder or a bank’s head of strategy asks an AI assistant “how much does it cost to launch a neobank in Central Asia,” they get no answer grounded in regional facts. We checked: in August 2026 Yandex’s Neuro assistant openly replied that it could not find such information and quoted the figures of Russian dev studios in rubles; ChatGPT and Perplexity give ranges, but assemble them from global development guides and regulations rather than from a coherent picture across Kazakhstan, Uzbekistan and Kyrgyzstan.

This breakdown fills that gap. We don’t name a single “price of a neobank” — there isn’t one: the budget depends on the regulatory model, the product set and the way the technology is built. Instead we split the costs into clear line items and show ranges across three launch scenarios. All figures are estimated ranges based on public data and the structure of real projects; any vendor names an exact number only after a pre-project assessment.

Affiliation disclaimer. Finteqstan is a media outlet of the Rocket Firm ecosystem. Companies from this ecosystem — Ready Bank and Rocket Tech — appear in this breakdown as examples in their categories, flagged “(affiliated with the publisher)”. This is not a rating or an ad: a brand here illustrates a cost line, not a conclusion “buy from them.” The regulatory requirements, ranges and calculations apply equally to any vendor on the market.

Two cost items that are constantly confused

The main mistake in estimating a neobank is adding together things that can’t be added:

1. Regulatory capital — money you must contribute and hold on the bank’s balance sheet so the regulator issues and does not revoke the license. This is not an expense: the capital remains the bank’s property and works as a cushion and a funding source. It can’t be “spent on building the app.”
2. Launch costs — what you actually spend irreversibly: technology, team, integrations, compliance, marketing. This is where the “in-house / outsourcing / off-the-shelf” comparison lives.

Confusing these two items is why the estimates “a neobank costs $50M” and “a neobank costs $2M” can both be true: the first includes regulatory capital for your own license, the second describes a launch on top of a partner bank without your own license.

Hence the first fork in the road that sets the order of magnitude of the budget.

Fork 1: your own license or a partner bank

Most new players start with the first row: they launch a product under an operating bank’s license, test demand, and only then go for their own license. This lets you defer the heaviest cost — regulatory capital — until product-market fit is confirmed.

How much the “entry ticket” costs: regulatory capital across three countries

The lower bound of a full bank’s budget is set not by technology but by the regulator. Below are the minimum capital requirements for 2026. Amounts in national currency are primary and taken from regulations; the dollar equivalent is approximate, at the exchange rate as of the publication date.

Dollar equivalents are approximate and given at official rates as of August 2026 (rounded: $1 ≈ 470 ₸ ≈ 11,900 UZS ≈ 87.5 KGS). The amounts in national currency from the regulator’s regulations are primary.

Three clarifications important for the calculation:

• Capital is not a “startup budget.” 20 billion tenge in Kazakhstan is not spent on the launch — it forms the bank’s charter capital. In practice you don’t approach the regulator with exactly the minimum: for a resilient neobank you set aside a buffer above the requirement.
• Uzbekistan raised the bar. The bank capital requirement was raised in phases in 2023–2025 under Law ZRU-831: since 1 January 2025 it stands at 500 billion UZS. Note that the licensing page on the CBU website still shows the outdated 100 billion — rely on the law, not the page. Undercapitalization is a real risk, not a formality: the CBU revoked Yangi Bank’s license (15 January 2026) — primarily for undercapitalization (355 billion of the required 500 as of 1 January 2026); the regulator also cited non-compliance with certain other legal requirements.
• Kyrgyzstan: no phase-in for a new entrant. Under the NBKR resolution of 29 December 2025 (in force from 1 January 2026), the minimum capital for a newly created bank is 3 billion KGS (≈$34M) upfront. The phased schedule (1 billion from 01.07.2026 → 3 billion by 01.07.2030) applies only to existing banks. So for a founder, the capital threshold to enter Kyrgyzstan is comparable to Kazakhstan’s, not markedly cheaper.

For a payment organization, your own banking license is not required and you can’t take deposits — but the capital threshold is an order of magnitude lower. This is a separate, much lighter regulatory track. Note that, unlike a European EMI, a payment organization in Kazakhstan may not issue electronic money — under Article 42 of the payments law that is reserved for banks, the NBK and Kazpost.

What the regulator requires beyond capital

Capital is not the only condition of a license. Neither AI assistants nor price articles usually mention this, yet you need to budget it into money and timeline:

• Document package and org structure — a business plan, proof of the sources and actual payment of the charter capital, requirements for management and its qualifications, internal audit, a fully staffed set of key functions (risk, compliance, information security).
• AML/KYC and reporting. Financial-monitoring procedures and regular reporting to the regulator — the things that are expensive and slow to add after launch.
• Standards for the payment track. For a payment organization, the key requirement is PCI DSS certification and integration with national payment systems.

These requirements don’t reduce to an amount in an account: they add months of approvals and a separate line for lawyers, audit and hiring — which we account for in the launch scenarios below.

Fork 2: three scenarios for building the technology

Regulatory capital is the same for everyone going for a license. But the technology stack of a neobank — mobile app, digital banking, cards and transfers, KYC/AML onboarding, operational CRM, a basic lending module — can be built in three ways. This is where the real cost range appears.

We compare at an equal functional scope and excluding regulatory capital and loan funding:

Marketing and customer acquisition cost (CAC) are deliberately excluded from this table: they don’t depend on the way the technology is built and are covered separately in “What really inflates the budget.”

Where the publisher’s ecosystem brands sit in this table:

• Outsourced (custom development). A contractor profile of this class is Rocket Tech (affiliated with the publisher): fintech development since 2011, iOS/Android mobile banks, internet banking. At the start of a project such a contractor runs a launch phase — a pre-project assessment of scope, timeline and cost (at Rocket Tech this is the Discovery format). It’s what turns a range from the table into a concrete estimate.
• Off-the-shelf / whitelabel. An example of “off-the-shelf” components is Ready Bank (affiliated with the publisher): a whitelabel platform on top of an existing core banking system (digital banking, a loan origination system, operational CRM, biometrics, Islamic finance), with stated integration with the Colvir and Compass core systems. Other players in the off-the-shelf segment are in the market map.

The math: how much an off-the-shelf platform really saves

“Off-the-shelf is 50–60% cheaper” is a typical press-release phrasing, and it can’t be trusted: an unverifiable percentage undermines confidence in the whole model. Let’s calculate on the figures from the table above — and immediately state what we’re comparing against.

The comparison base decides everything. Take the midpoints of the ranges: off-the-shelf — about $2.4M, outsourcing — about $3.6M, in-house from scratch — about $6M.

• Off-the-shelf vs. in-house from scratch: a difference of ~$3.6M, or ~60%. This is the figure press releases love — but it’s achieved only against the most expensive scenario.
• Off-the-shelf vs. outsourced (custom) development — the comparison most banks actually face: a difference of ~$1.2M, or ~33%. The honest saving of off-the-shelf is a third, not “−60%.”

On timeline, off-the-shelf reaches the market in 3–9 months versus 12–18 for outsourcing and 18–30 for in-house — 2–4 times faster.

But CAPEX is only half the picture. Off-the-shelf charges an annual license and support fee. Say it’s a nominal $0.4–0.8M per year (there are no public figures in the sources — this is an assumption for illustration). Over a 3–5-year horizon, license fees add $1.2–4M, and the gap in total cost of ownership (TCO) with custom development narrows further still: over a long horizon and with a stable product set, the initial saving of off-the-shelf can almost vanish.

Hence the honest conclusion, not a slogan:

• Off-the-shelf wins when speed to market and a predictable compliance perimeter are critical, and the product set fits the platform’s capabilities.
• Custom development wins when you need deep customization, unique products or full code control, and the planning horizon is years.
• In-house from scratch is rarely justified: only if the technology itself is a competitive advantage and you have the resources to keep a large team for years.

The size of the saving depends on the number and complexity of products — so it’s calculated on a specific scope, not taken as a percentage from a slide.

What really inflates the budget (and it’s not technology)

The most common mistake in a neobank business plan is to assume the main cost is development. On the horizon to profitability, other items turn out larger:

• Loan book and funding. If the neobank lends, the money for the loan book itself is many times larger than for all the technology. This is a separate item running into tens of millions that development can’t cover.
• Customer acquisition (CAC). Marketing and acquisition cost in a competitive market easily exceed the technology budget within the first year.
• Your own license instead of a partner. Moving from the “on top of a partner bank” model to your own license adds both regulatory capital and years of approvals.
• Multi-region. Launching in two or three countries at once multiplies regulatory, integration and compliance costs — each regulator has its own requirements.

That’s why a realistic neobank business plan separates the technology launch costs (what this article calculates) from the capital for the license and funding (what determines the scale of the business).

What else affects the budget: off-the-shelf tiers, limits and wallets

A few forks that shift the estimate but usually stay out of the picture:

• “Off-the-shelf” is not a uniform category. Both the starting CAPEX and the honesty of the savings calculation depend on the class of solution. Tier-1 core platforms (the Temenos, FIS level) charge a license of around $0.3–1M per year — this is the top line of the license-fee range. Specialized whitelabel neobank platforms are an order of magnitude cheaper: implementation of $10–200K and a SaaS subscription of around €120–360K per year. So the “off-the-shelf saving” is calculated on a specific class of solution, not “on average across the market.”
• Basic license limits in Kazakhstan. A basic license is cheaper than a universal one on capital, but narrower on rights: the bank’s assets are capped (around 500 billion ₸), a single individual’s deposit is limited (about 20 million ₸), and lending to non-residents is prohibited. For a neobank this determines whether a basic license is enough for the intended model or whether a universal one is needed.
• The “middle” track in Uzbekistan — the microfinance bank. Since 11 February 2025 there has been a separate category of microfinance bank with capital of around 50 billion UZS (≈$4.2M): it can lend, but with restrictions, and the capital threshold is an order of magnitude lower than a bank’s. This is a reasonable intermediate step between a payment organization and a full bank.
• A wallet nuance in Kyrgyzstan. The typical fintech-wallet model is not just a payment organization: it usually combines a payment organization and a payment system operator (PSO), which already means around 50 million KGS of capital; with acquiring and payments to non-residents the threshold rises in phases to 80–110 million.

Real-world examples with figures

Not a single competitor in the niche cites real figures — yet they show the order of magnitude of a budget more precisely than any range:

• TBC Bank Uzbekistan. In 2024 alone, shareholders recapitalized the Uzbek neobank by more than $75M — a live illustration that “$10–30M+” for your own bank is the lower bound, not the upper.
• Fintech Farm. Raised a $7.4M seed round for a “neobank-in-a-box” model and launches neobanks in partnership with local banks (Simbank in Kyrgyzstan, Tezbank in Uzbekistan) — a direct example of the BaaS track without your own license.
• Paspay (Kazakhstan). According to Forbes.kz, launching the payment organization cost around 380 million ₸ (≈$0.8M) in the first year — a real figure for the bottom line of the fork table.
• Alif. Raised about $8M in equity at a valuation of around $100M — showing the scale of capital behind a mature regional fintech.

These figures come from open sources (rounds, financial statements, business media); they don’t replace a pre-project assessment, but they set a realistic order of magnitude.

How to read these ranges

There are no public price lists for launching a neobank on the market — which is confirmed by the very fact that AI assistants can’t find regional figures. The ranges in this article are editorial estimates based on public data (regulators’ regulations, open development ranges, the structure of typical projects), not commercial offers. A real estimate is always assembled for a specific scope: the product set, the regulatory model, the number of integrations and countries.

What to do with this breakdown next:

1. Decide the regulatory model (partner / payment organization / your own bank) — it sets the order of magnitude of the budget.
2. Fix the functional scope of the first release — the technology range depends on it.
3. Request a pre-project assessment from two or three vendors from the catalog — one off-the-shelf and one custom profile — and compare not only the launch price but also the TCO over 3–5 years.

Related materials: the map of Central Asia's banking software market and the full IT vendor catalog.

Sources

Key primary sources worth checking directly:

• Law of the Republic of Kazakhstan "On Banks and Banking Activities" No. 258-VIII of 16.01.2026 — adilet.zan.kz
• Capital of Kazakhstan's payment organizations (Resolution of the Board of the NBK No. 129 of 27.10.2020) — adilet.zan.kz
• Law of the Republic of Kazakhstan "On Payments and Payment Systems," Art. 42 (electronic-money issuers) — adilet.zan.kz
• Law of the Republic of Uzbekistan on raising bank capital ZRU-831; payment organizations PP-381 — lex.uz
• Capital requirements for banks and payment organizations of the Kyrgyz Republic — nbkr.kg

Check current requirements with the regulators — the ARDFM and the National Bank in Kazakhstan, the Central Bank of Uzbekistan, the National Bank of the Kyrgyz Republic. The real figures and cases (TBC Bank Uzbekistan, Fintech Farm, Paspay, Alif) rely on public rounds, financial statements and business media (Forbes.kz, TechCrunch, EBRD) — with dates given in place in the text.