The retail prop firm industry has a tech stack problem that no operator wants to talk about publicly. The dominant stack (MetaTrader 5 plus a thin CRM layer plus custom-built risk and compliance) was acceptable in 2019 when license fees were reasonable and customer expectations were modest. In 2026 it's a margin killer. The operators still running it are increasingly losing the unit-economics competition to operators on modern purpose-built infrastructure.
This article runs the actual numbers on a mid-sized prop firm operating on a legacy stack versus modern infrastructure. Where we cite specific costs, these are based on observed pricing from operators in our B2B partner network and from public information about MetaQuotes' license schedules.
The four cost categories of a legacy stack
A prop firm running on MetaTrader 5 has costs in four layers:
Layer 1: MetaQuotes licensing
The MT5 server license has shifted dramatically over the last four years. The 2019-era licensing was approximately $20K-$25K annually for a standard prop firm setup with a few thousand accounts. Current published pricing for the same setup is approximately $55K-$75K annually, with additional per-account fees for accounts beyond certain thresholds.
For a 5,000-active-trader prop firm in 2026, the all-in MetaQuotes licensing cost is typically $80K-$120K per year, depending on configuration. This is just for the right to use the platform; it doesn't include hosting, customization, or any of the layers below.
Layer 2: White-label hosting and infrastructure
Most prop firms don't host MT5 themselves. They use one of the white-label providers (B2Broker, Match-Trader, others) who manage the server infrastructure, liquidity bridge, and basic operational layer. White-label costs typically run $5K-$15K per month for a mid-sized prop firm, plus per-trade routing fees.
Annual cost: $60K-$180K per year.
Layer 3: Customization for prop firm workflows
MT5 is a retail brokerage platform, not a prop firm platform. The prop-firm-specific workflows. Multi-stage evaluation challenges, profit-share accounting, trader rule violation detection, payout processing. All have to be built as customizations on top. Most prop firms either build these in-house (expensive, slow) or buy them from one of the prop-firm-specific software vendors (also expensive, but at least not slow).
The reality at most operators: initial customization development costs $80K-$200K for a functional prop firm layer. Ongoing maintenance and feature development runs $40K-$80K per year minimum, with significant spikes when MetaQuotes ships major platform updates that break existing customizations.
This is the most underestimated cost. Operators tend to budget for the initial build but not for the ongoing maintenance overhead of customizations that break every time MetaQuotes ships a major update.
Layer 4: CRM, risk, payments, compliance
Everything else. CRM customization for prop firm workflows: $30K-$60K/year. Risk engine (if built rather than bought): meaningful engineering investment. Payment processing and reconciliation: 1-2% of revenue in fees and operational overhead. Compliance and KYC tooling: $20K-$50K/year.
The total cost picture
Summing the four layers for a representative 5,000-active-trader prop firm running on a legacy MT5 stack:
| Layer | Annual cost (low end) | Annual cost (high end) |
|---|---|---|
| MetaQuotes licensing | $80,000 | $120,000 |
| White-label hosting | $60,000 | $180,000 |
| Customization (initial amortized + ongoing) | $120,000 | $280,000 |
| CRM / risk / payments / compliance | $80,000 | $160,000 |
| Total platform-layer costs | $340,000 | $740,000 |
The same operation on modern purpose-built infrastructure (Arizet's stack, for full disclosure on the comparison source):
| Layer | Annual cost |
|---|---|
| A-Trader for Prop Firms (Growth tier) | $120,000 |
| Prop-Tech CRM (Growth tier) | $180,000 |
| Prop Risk (Discovery + value-share, year 1) | $96,000 base + value-share |
| Total platform-layer costs | $396,000 |
At first glance these look similar. The actual difference shows up in two places: operational efficiency and quality of features.
The hidden differences
Beyond raw cost, three structural differences favor modern infrastructure:
1. Time to launch new products. On legacy MT5, launching a new evaluation product (different rules, different account size, different profit split) typically takes 4-8 weeks of customization work. On modern infrastructure, the same change is a configuration update. Minutes to hours. For an operator running 6-8 evaluation products and adjusting them frequently based on market response, this is the difference between rapid iteration and slow death.
2. Quality of risk detection. First-generation risk engines detect rule violations (hard drawdown breaches). Modern engines detect pattern violations (coordinated trading clusters, gaming behavior, suspicious statistical signatures). The savings from catching coordinated payout fraud rings (which are increasingly common in this space) can be in the millions for a mid-sized operator. Prop Risk has documented $20M+ in saved leakage across partner firms over three years.
3. Customer experience. Modern platforms offer experiences (mobile parity, instant deposits, real-time analytics, social features) that drive better trader retention. On legacy MT5, retention rates are typically 35-45% at month 12. On modern stacks with engagement features, retention typically runs 55-70%. Over a multi-year period, the LTV difference compounds dramatically.
The compounding margin problem
Here's where it gets uncomfortable. The legacy stack costs are increasing while modern stack costs are decreasing:
Legacy trajectory:
- MetaQuotes license fees have increased every year since 2021. Industry expectation is for further increases.
- White-label provider fees have increased as the providers themselves face higher costs.
- Customization debt compounds. Every new MT5 update requires re-validation and often re-implementation of custom features.
- Regulatory compliance overhead is increasing, and legacy systems handle this poorly.
Modern trajectory:
- Cloud infrastructure costs decline ~10-15% per year on average.
- AI-driven workflows reduce operational overhead per trader meaningfully.
- Multi-tenant SaaS economics allow vendors to lower pricing as customer base grows.
- Modern stacks handle compliance updates centrally rather than per-customer.
The gap between the two cost trajectories widens every year. A firm that's break-even on its tech costs today on legacy infrastructure will be unprofitable on tech costs in 18 months. A firm that's investing in modern infrastructure today is locking in unit economic advantages that compound.
When does migration become obvious?
Migration is hard. We won't pretend otherwise. Moving a live prop firm from MT5 to modern infrastructure is typically a 6-12 week project for a mid-sized operator, with real risks (downtime, trader confusion, data migration errors). The question isn't whether migration is hard. It's whether the cost of not migrating exceeds the cost of migration.
The threshold we observe across operators: when annual platform-layer costs exceed 12-15% of revenue, migration becomes mathematically obvious. Most legacy MT5 operators are currently at 15-22%. Modern operators are at 6-10%. The gap is meaningful.
The honest math
For a $5M ARR prop firm currently spending $750K/year on a legacy stack, migrating to a modern stack typically saves $300K-$450K/year in direct costs plus $200K-$500K/year in productivity gains (faster product launches, better retention, better risk detection). The migration project costs $50K-$150K and takes a quarter to complete. Payback under 6 months. The math is decisive once you run it honestly.
The reason most operators don't migrate isn't economic. It's organizational. Switching costs feel large in the moment because they're concentrated, while the benefits compound over time. The operators who've migrated successfully in 2024-2026 are now seeing dramatically better unit economics than their legacy-stack peers. The operators who keep waiting will face migration with weaker margins and less flexibility.
The math says do it now. We say so partly because we benefit from new partnerships, and partly because we've watched too many operators hit a wall they could have avoided by moving 18 months earlier.