Priced like infrastructure, not like a meter.
Sovereign deployments are scoped to workloads, sites and service levels. Every engagement starts with a sizing exercise rather than a price list.
Pilot
From one node
Prove a use case on dedicated hardware in a single site.
- Single GPU node or small cluster
- Two production applications
- Platform governance included
- Business-hours in-region support
- Fixed-scope deployment sprint
Enterprise
Multi-rack estate
A production AI estate for an organisation and its regulators.
- Multi-rack GPU cluster
- Unlimited applications and teams
- Hybrid sovereign cloud bursting
- 24/7 support with contractual SLAs
- Disaster recovery and DR testing
- Quarterly capacity planning
National
Multi-site programme
Country-scale infrastructure across core, regional and edge sites.
- Multi-site, multi-region topology
- Air-gapped and classified enclaves
- Dedicated engineering team on site
- Skills transfer and accreditation
- Sovereign model programme
- Long-term operating agreement
Commercial questions, answered plainly.
Do you charge per token?
No. You pay for infrastructure and platform, not for the intelligence it produces. Usage metering exists for internal chargeback, not for our invoice.
Can we buy the hardware ourselves?
Yes. We can deploy onto hardware you procure, or deliver the full stack as one programme.
What happens if we leave?
You keep the hardware, the open-format models and the data. Portability is contractual, not a promise.
Is there a cloud option?
Yes, in sovereign in-country regions, typically combined with on-premise inference in a hybrid design.
Get a sizing and cost model for your estate.
Talk to our sovereign infrastructure team about a deployment scoped to your jurisdiction, workloads and compliance obligations.