Core Banking TCO: Real Costs for SaaS, Self-Hosted & Hybrid [2026] - CoreFi
CoreFi · 12 min read
Every core banking vendor has a slide that says "Lower TCO." None of them show you the actual numbers.
We've worked with financial institutions that spent 3x their initial budget on platform costs because the real expenses were buried in licensing terms, integration fees, and operational overhead nobody mentioned during the sales cycle.
This article lays out the honest math. Real cost ranges. Real hidden fees. Real 5-year projections across deployment models.
Why TCO Is So Hard to Calculate
Core banking TCO is deliberately opaque because vendors benefit from complexity. Here's what makes it hard:
- Pricing models vary wildly: per-account, per-transaction, per-API-call, flat license, revenue share, or combinations
- Integration costs are externalized: the vendor quotes the platform, but you pay separately for every connection
- Scaling costs are non-linear: cheap at 10K accounts, expensive at 500K
- Exit costs are hidden: data extraction, format conversion, parallel running during migration
- Compliance costs are yours: the vendor provides "tools," but implementation is your problem
The Three Deployment Models
Model 1: Full SaaS (Cloud-Hosted, Vendor-Managed)
Examples: Mambu, Tuum, Thought Machine (Vault as SaaS)
How pricing typically works:
- Monthly platform fee based on number of active accounts
- Per-transaction fees for payments, transfers, and card operations
- API call volume tiers
- Additional modules priced separately (lending, cards, compliance)
Year 1 Cost Breakdown (100K accounts, mid-market):
| Cost Category | Range | Notes |
|---|---|---|
| Platform license | β¬200-500K/yr | Per-account pricing varies 10x between vendors |
| Implementation & onboarding | β¬150-400K | Vendor professional services + internal team |
| Integration (payment rails, KYC, cards) | β¬100-300K | 5-10 integrations at β¬20-30K each |
| Data migration | β¬50-150K | Schema mapping, validation, parallel running |
| Internal team (project) | β¬200-400K | 3-5 FTEs for 12 months |
| Testing & UAT | β¬50-100K | Performance, security, compliance testing |
| Training | β¬20-50K | Technical and business user training |
| Year 1 Total | β¬770K-1.9M |
Ongoing Annual Cost (Year 2+):
| Cost Category | Range | Notes |
|---|---|---|
| Platform license | β¬200-500K/yr | Grows with account volume |
| Transaction fees | β¬50-200K/yr | Depends on transaction volume |
| Integration maintenance | β¬30-80K/yr | API version updates, new integrations |
| Internal team (BAU) | β¬150-300K/yr | 2-3 FTEs for operations and development |
| Annual Ongoing | β¬430K-1.1M |
5-Year TCO: β¬2.5M-6.3M
Model 2: Self-Hosted (On-Premises or Private Cloud)
Examples: Temenos (T24/Transact), Finastra, FIS, traditional on-prem vendors
How pricing typically works:
- Upfront perpetual license fee (large)
- Annual maintenance fee (15-22% of license)
- Infrastructure costs entirely on you
- Professional services for every customization
Year 1 Cost Breakdown (100K accounts, mid-market):
| Cost Category | Range | Notes |
|---|---|---|
| Perpetual license | β¬500K-2M | One-time, but often requires renegotiation at scale |
| Annual maintenance | β¬75-440K/yr | 15-22% of license fee |
| Implementation | β¬300-800K | Long implementation cycles (12-24 months) |
| Infrastructure (servers, DB, network) | β¬100-300K/yr | Private cloud or on-prem |
| Integration | β¬150-400K | Typically more complex with legacy systems |
| Data migration | β¬100-250K | Legacy data formats require extensive transformation |
| Internal team (project) | β¬400-700K | 5-8 FTEs for 18-24 months |
| Year 1 Total | β¬1.6M-4.9M |
Ongoing Annual Cost (Year 2+):
| Cost Category | Range | Notes |
|---|---|---|
| Maintenance fee | β¬75-440K/yr | Non-negotiable, required for updates |
| Infrastructure | β¬100-300K/yr | Servers, DBAs, monitoring, DR |
| Internal team (BAU) | β¬300-500K/yr | 4-6 FTEs (DBA, infra, app support, dev) |
| Customization & upgrades | β¬100-300K/yr | Every major version requires project |
| Annual Ongoing | β¬575K-1.5M |
5-Year TCO: β¬3.9M-11M
Model 3: Hybrid (SaaS Platform, Self-Managed Extensions)
Examples: CoreFi, some Mambu deployments, Skaleet
How pricing typically works:
- SaaS platform fee (lower than full SaaS due to shared responsibility)
- Infrastructure for custom modules and extensions is yours
- API-first integration reduces coupling costs
Year 1 Cost Breakdown (100K accounts, mid-market):
| Cost Category | Range | Notes |
|---|---|---|
| Platform license | β¬120-300K/yr | Lower per-account cost with self-managed extensions |
| Implementation | β¬100-250K | Faster due to API-first, modular architecture |
| Integration | β¬80-200K | Standardized APIs reduce per-integration cost |
| Custom module development | β¬50-150K | Build differentiated features on top of core |
| Data migration | β¬40-100K | Modern formats, automated tooling |
| Internal team (project) | β¬150-300K | 2-4 FTEs for 6-9 months |
| Year 1 Total | β¬540K-1.3M |
Ongoing Annual Cost (Year 2+):
| Cost Category | Range | Notes |
|---|---|---|
| Platform license | β¬120-300K/yr | Predictable scaling |
| Extension hosting | β¬20-60K/yr | K8s cluster for custom modules |
| Internal team (BAU) | β¬100-250K/yr | 1-3 FTEs |
| Annual Ongoing | β¬240-610K |
5-Year TCO: β¬1.5M-3.7M
The Hidden Costs Nobody Mentions
1. Contract Lock-In Penalties
Most vendors require 3-5 year commitments. Early termination fees range from 50-100% of remaining contract value. A 3-year contract at β¬300K/year means β¬450-900K if you need to switch after year one.
Ask before signing: What's the exact early termination fee? Is there a mutual termination clause?
2. Data Extraction Costs
When you leave a platform, getting your data out is rarely free or easy. Expect:
- β¬50-200K for data export in a usable format
- 3-6 months of parallel running (paying both old and new vendor)
- Data format conversion engineering effort
Ask before signing: Do you provide data export in standard formats (ISO 20022, CSV)? Is there a documented exit process?
3. API Rate Limiting
SaaS vendors often include API call limits in base pricing. Exceeding them triggers overage fees that can double your monthly bill.
Example: 1M API calls/month included, β¬0.001 per additional call. A high-traffic deployment doing 5M calls/month pays β¬4,000/month in overages β β¬48K/year unbudgeted.
Ask before signing: What are the API rate limits? What are overage costs? Can we negotiate unlimited API access?
4. Compliance Module Upsells
The base platform rarely includes compliance tools. AML screening, transaction monitoring, regulatory reporting, and audit trails are typically separate modules with separate pricing.
Common add-ons and costs:
- AML/KYC module: β¬30-80K/year
- Transaction monitoring: β¬40-100K/year
- Regulatory reporting: β¬20-50K/year
- Audit trail & compliance dashboard: β¬15-40K/year
- Total compliance add-ons: β¬105-270K/year
5. Version Upgrade Projects
Self-hosted platforms require major version upgrades every 2-3 years. Each upgrade is essentially a mini-migration project:
- 3-6 months duration
- 2-4 dedicated engineers
- Regression testing across all integrations
- Cost per upgrade: β¬100-400K
SaaS platforms handle this automatically β but may introduce breaking changes that require your integration code to be updated.
6. Expertise Tax
Specialized core banking platforms require specialized (expensive) engineers. Temenos developers command β¬600-1,000/day. Thought Machine's Smart Contracts language requires dedicated training.
The math: If your platform requires 2 specialized engineers at β¬150K/year above market rate, that's β¬300K/year in expertise premium β rarely accounted for in TCO calculations.
5-Year TCO Comparison Summary
For a mid-market fintech with 100K accounts growing to 500K:
| Model | Year 1 | Year 2-5 | 5-Year TCO | TCO/Account/Year |
|---|---|---|---|---|
| Full SaaS | β¬1.3M | β¬3M | β¬4.3M | β¬2.90 |
| Self-Hosted | β¬3.2M | β¬4.2M | β¬7.4M | β¬4.90 |
| Hybrid | β¬900K | β¬1.7M | β¬2.6M | β¬1.70 |
Mid-range estimates. Actual costs vary significantly by vendor, geography, and complexity.
How to Run Your Own TCO Analysis
Step 1: Map your requirements β List every capability you need (accounts, payments, lending, cards, compliance, reporting)
Step 2: Get all-in pricing β Don't accept "starting from" quotes. Get written pricing for your specific volume, including all modules, integrations, and compliance tools
Step 3: Add internal costs β Engineering team, project management, testing, training. Use your actual loaded cost per engineer, not vendor estimates
Step 4: Model growth scenarios β Price at current volume AND projected 3-year volume. Many vendors have aggressive scaling tiers
Step 5: Price the exit β Calculate what it would cost to leave after year 2. Include data extraction, parallel running, and new vendor onboarding
Step 6: Add 30% contingency β Every core banking project we've seen has exceeded initial estimates. Budget accordingly.
The CoreFi Approach
CoreFi uses a hybrid model designed to minimize TCO:
- Modular pricing: Pay only for the modules you use (core banking, lending, tokenization, custody)
- API-first: Standard REST APIs reduce integration costs by 40-60% vs proprietary interfaces
- No per-transaction fees: Flat pricing that doesn't penalize growth
- Built-in compliance: MiCA, DORA, and AML capabilities included in base platform
- Open data: Standard export formats, documented migration tooling, no lock-in penalties
We publish our pricing because we believe transparency builds trust. If a vendor won't give you all-in pricing in writing before you sign, that tells you everything you need to know about their TCO.
Want a personalized TCO analysis for your institution? CoreFi provides transparent, modular pricing with no hidden fees. Request a detailed quote β we'll show you the real numbers.