Pricing & deployment models
Every production engagement begins with measured candidate ranking, path analysis, and a runtime recommendation. The approved configuration is then deployed in Managed Cloud, Customer Cloud, or Self-Hosted mode.
Benchmark sprint
The benchmark sprint is the common entry point across all deployment models.
Choose your deployment model
Where the runtime lives determines billing, account ownership, and control boundaries.
Deploy on HFTCloud-operated infrastructure and launch through the HFTCloud control panel. HFTCloud provisions, tunes, and operates the runtime so the team gets the shortest path from benchmark to production with one operational counterpart and one invoice.
Deploy into your own AWS account while keeping HFTCloud placement logic, orchestration, and tuned deployment workflow. You keep ownership of cloud spend, access policy, and third-party infrastructure contracts. HFTCloud is billed separately.
Install the HFTCloud package inside your own environment and operate within your internal security boundary. This is the highest-control model for teams that require self-hosted deployment, tighter review, and stricter change management.
Comparison matrix
Use the matrix below to choose the operating model that fits your desk.
| Item | Managed Cloud | Customer Cloud (BYOC) | Self-Hosted |
|---|---|---|---|
| Where the runtime lives | HFTCloud-operated environment | Customer AWS account | Customer environment / internal perimeter |
| Who owns the cloud account | HFTCloud | Customer | Customer |
| Connectivity contracts | Included through HFTCloud where applicable | Customer-managed where applicable | Customer-managed |
| Billing model | Setup fee + monthly fee paid to HFTCloud | Setup fee + HFTCloud monthly fee, plus direct infrastructure bills | Setup fee + annual software/support agreement, plus direct infrastructure bills |
| How the team launches | Through HFTCloud control panel | Through HFTCloud control panel with customer AWS access | Download package and deploy inside your environment |
| Time to first run | Fastest | Fast after access setup | Slowest, review-led |
| Benchmark eligibility | Default starting path | Available after qualification | Limited, technical review first |
| NDA timing | Later if needed | Earlier for some scopes | Early, usually before package review |
| Customer operational burden | Lowest | Medium | Highest |
| Control and isolation | Standard | High | Highest |
Support & engineering
This separates infrastructure ownership from engineering scope and makes the commercial boundaries explicit.
Onboarding into the selected deployment model, product support, platform updates, and normal rollout guidance.
Faster response windows, deeper operational coordination, and more structured change planning.
Custom integrations, environment-specific tuning, and advanced performance work beyond standard product support.
Trading stack options
HFTCloud pricing covers the infrastructure and deployment layer. Application foundations, the HFTKernel runtime, and market data scope are qualified separately when required.
Multi-asset, multi-venue trading workloads built on a Rust-native engine with Python strategy control.
Exchange adapters, data collection, and execution workflows built on one consistent crypto trading API.
Strategy research, backtesting, optimization, and repeatable live deployments for Python-native teams.
Latency-oriented Linux runtime, topology-aware tuning, and release-aligned cpujitter and memjitter validation.
Validated feeds, timing provenance, order-book reconstruction, replay, and production market-data workflows.
FAQ
Qualified teams can start with a 5-business-day benchmark sprint at no charge. Production rollout moves to an annual commercial agreement once the scope is validated.
No. The benchmark can run alongside the current setup so teams can compare the current path against feasible candidates before broader changes.
Yes. Many teams start in Managed Cloud to validate quickly, then move to Customer Cloud or Self-Hosted when procurement, security, or scale requires it.
Extended support, custom integrations, and non-standard engineering are scoped separately from the deployment model.
Higher-control models and venue-sensitive rollout details may require an NDA earlier in the process.
Start with a benchmark sprint
Begin with one venue, one workload profile, and one measurement objective. Then choose the deployment model and support scope that fit your team.