What It Costs a 300,000-Customer Bank to Run an Open Banking API Platform
A hypothetical five-year model puts a vendor platform near $3.9M and a gateway-plus-build near $6.0M for a 300,000-customer bank. Here's where the money goes, and when building still wins.
The buy-or-build question a mid-sized bank actually faces
If you run technology at a bank with around 300,000 retail customers, open banking rarely arrives as a strategy debate. It arrives as an obligation with a deadline. Your regulator expects third-party providers (TPPs) to reach customer accounts through secure APIs, and you must choose: license a vendor open banking platform, or put an API gateway in front of your core and build everything else yourself.
Most published open banking API platform pricing covers only the license line, and the license is rarely the biggest number. The real five-year cost sits in consent handling, TPP onboarding, fallback arrangements, availability reporting and the engineers who keep up with standards that change regularly. This piece builds a clearly labelled hypothetical cost model around those lines.
Hypothetical bank used throughout: 300,000 retail customers, one core banking system, mobile and web channels, and an obligation to support account information and payment initiation under a PSD2-style regime. We assume a blended loaded engineering cost of about $140,000 per year, plausible for parts of Europe and the Gulf. It runs higher in London, Zurich or Sydney and much lower in India or much of Southeast Asia, so rescale the labor lines for your market.
All figures are illustrative, typical 2026 market ranges that vary heavily by contract, region and scope. Confirm every number with vendors before you budget.
Five-year cost model: vendor platform versus gateway plus in-house build
The vendor path assumes a specialist account-servicing platform, such as Ozone API, or an open banking module from a core vendor like Temenos or Finastra, deployed in the bank's cloud tenancy or as managed SaaS. The build path assumes an API management product (Google Apigee, Kong, MuleSoft, IBM API Connect, Axway or WSO2, some with open banking accelerators) plus in-house development of everything specific to open banking.
| Cost line (USD thousands) | Vendor: Year 1 | Vendor: Years 2-5, per year | Build: Year 1 | Build: Years 2-5, per year |
|---|---|---|---|---|
| Implementation or build labor (incl. integrator) | 400 | 0 | 1,090 | 0 |
| Platform subscription or gateway license | 250 | 250 | 150 | 150 |
| Consent management | Included | Included | 100 | 100 |
| Internal run team | 300 | 300 | In build labor | 560 |
| Hosting and monitoring | 40 | 40 | 100 | 100 |
| Security testing and conformance | 80 | 60 | 100 | 80 |
| Standards and regulatory change | 0 | 50 | 0 | 120 |
| Total | 1,070 | 700 | 1,540 | 1,110 |
| Five-year total | About 3,870 | About 5,980 |
In this example the build path costs roughly 1.5 times as much over five years. Two assumptions drive most of the gap. The build team needs about six engineers in year one and four afterward, against two or three on the vendor side. And every standards release, whether a new Berlin Group version, a UK Open Banking update or a CDR Consumer Data Standards change, lands on your backlog instead of the vendor's roadmap.
The gap narrows sharply if the bank already runs an enterprise gateway and an API team. Take the gateway license out and cut the run team to two or three incremental heads, and the build path lands close to the vendor path. That's why API management cost should be judged at the margin, not as a standalone purchase.
The cost lines that blow up budgets
Consent management
Consent management is more than a database table. It has to capture granular permissions, enforce expiry and re-authentication cycles, support revocation from both the bank's channels and the TPP side, and give customers a dashboard. Builders either license a standalone consent product or spend months writing one. Under-scoped consent is a common reason in-house projects overrun.
TPP onboarding
Every TPP needs identity verification against a trusted source. In the EU that means validating eIDAS qualified certificates and checking authorization status. In the UK it means integrating with the Open Banking directory. In Australia it means the CDR Register. Dynamic client registration, certificate rotation and handling revoked TPPs are ongoing operational work, not a one-off build.
Fallback interfaces
Under PSD2, a dedicated interface must generally come with a fallback mechanism unless the national authority grants an exemption. Earning one means meeting performance, testing and usage conditions. Running a fallback can add an indicative $50,000-$150,000 a year in engineering and monitoring, so plan for the exemption from day one. Ask vendors how many of their clients have obtained one.
Availability and performance reporting
Regulators expect uptime and response-time statistics for the API, compared with your own customer channels. Vendors usually ship this as a reporting module. Builders often find gateway logs don't produce the required metrics without custom work, which can plausibly run $40,000-$100,000.
How the bill changes by jurisdiction
The same 300,000-customer bank faces different scope, and so different cost, depending on where it operates. Regimes are converging in concept but not in technical detail.
| Regime | What the bank must provide | Cost implication |
|---|---|---|
| UK | Account access under the payment services rules, typically on the UK Open Banking standard; directory integration; regular availability reporting | Mature vendor market; commercial variable recurring payments may add payment scope |
| EU | PSD2 dedicated interface, often Berlin Group NextGenPSD2 or STET; fallback or exemption; eIDAS certificate checks | Proposed PSD3 and Payment Services Regulation changes may alter fallback and dashboard rules; proposed open finance rules could widen data scope |
| Saudi Arabia | APIs under SAMA's Open Banking Framework, phased from account information to payment initiation | Data residency and cloud rules can rule out offshore SaaS, which raises hosting cost |
| Australia (CDR) | Data holder APIs to the Consumer Data Standards; performance requirements; metrics reporting; consumer dashboard | Strict uptime and performance thresholds push up hosting and on-call cost |
| India (Account Aggregator) | Financial information provider integration with licensed account aggregators using ReBIT specifications | Consent is managed by the aggregator, so the bank's consent build is far smaller; total cost is often a fraction of the EU model |
For a bank in India, the model above overstates costs considerably. A financial information provider integration can often be delivered through an established technology service provider for well under the EU-style figures. For a Saudi bank, the vendor path works only if the vendor can deploy in-kingdom. Check that first.
Red flags in vendor quotes and a decision rule
Watch for these when comparing proposals:
- Per-call pricing with no cap. TPP traffic can climb sharply once aggregators poll accounts several times a day. Insist on tiered bands or an annual ceiling.
- Standards upgrades billed as change requests. Mandatory standard versions belong in the subscription.
- Consent and reporting sold as add-ons. If either costs extra, re-run the comparison, because it erases much of the vendor advantage.
- No reference clients in your jurisdiction. A platform proven only on the UK standard may need heavy work for Berlin Group, CDR or SAMA specifications.
- Vague exit terms. Confirm you can export consent records and TPP registrations in a usable format.
The decision rule: if your bank doesn't already run an enterprise API gateway with a dedicated team of at least three engineers, buy a vendor platform. Your obligations are compliance-driven, and a vendor absorbs standards churn more cheaply than you can. If you already have that gateway and team, and you see open APIs as the base for commercial products such as premium APIs, embedded finance or business banking data feeds, building can pay off. Model it at marginal cost, not full cost.
Your next step: take the table above, swap in your local labor rates and actual gateway position, and send vendors a scope list covering all seven cost lines. Ask them to price each line separately, and have compliance counsel confirm the scope of your local obligations before you sign.
Common questions
What does an open banking API platform typically cost a mid-sized bank each year?
As a typical 2026 market range, vendor subscriptions for a bank with a few hundred thousand customers often fall between $150,000 and $400,000 a year, with implementation of $250,000 to $600,000 on top. Contracts vary widely by region, deployment model and traffic, so treat these as starting points and confirm with vendors.
Is building on our existing API gateway really cheaper?
Only if the gateway and API team already exist and are funded for other uses. If you must buy the gateway and hire a team specifically for open banking, the in-house path usually costs more over five years, because consent, TPP onboarding, reporting and standards updates all become your responsibility.
Do we need a fallback interface in addition to our dedicated API?
Under PSD2 in the EU, a fallback mechanism is generally required unless your national authority grants an exemption based on the dedicated interface's performance and usage. UK rules differ in detail, and proposed changes under PSD3 and the Payment Services Regulation may alter the EU position, so confirm with your regulator and legal advisers.
Should consent management be bought separately or come with the platform?
For most mid-sized banks, consent management should come inside the open banking platform, because it is tightly coupled to authorization flows and customer dashboards. A standalone product makes more sense if you need consent handling across several regimes or for non-open-banking data sharing.
How different is the cost for a bank in India's Account Aggregator framework?
Usually much lower. Banks acting as financial information providers integrate with licensed account aggregators that manage consent, and many use established technology service providers for the integration. The EU-style figures in this model overstate Indian costs considerably.
What pricing terms should we negotiate hardest?
Focus on traffic-based fees, coverage of mandatory standards upgrades, and whether consent and availability reporting modules are included. Also negotiate export rights for consent records and TPP registrations so you can switch vendors without rebuilding your compliance history.
Written by the Wealth CRM Direct editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.