Sovereign data centers and GPU clusters — the compute foundation for nations and operators.
Soika Stack runs on the GPU estate we source, deliver, install and commission for you — from a single NVIDIA HGX B300 node to GB300 NVL72 racks and next-generation Vera Rubin systems — inside your data centre or sovereign cloud, then operated with your partner.
HGX B300, GB300 NVL72 and Vera Rubin — delivered and deployed as a service.
NVIDIA HGX B300
Eight Blackwell Ultra GPUs on a single HGX baseboard — the building block for dense inference and training servers, delivered in air- or liquid-cooled configurations from our OEM partners.
NVIDIA GB300 NVL72
72 Blackwell Ultra GPUs and 36 Grace CPUs in one liquid-cooled rack, NVLink-connected to behave as a single accelerator for frontier-scale inference and reasoning workloads.
NVIDIA Vera Rubin
NVIDIA's next-generation Rubin GPU and Vera CPU platform. We plan capacity and reserve allocation ahead of availability, so your roadmap is not gated by the supply queue.
- Delivery, installation & commissioning
- Validated fabric, power & cooling
- Operated as a service with your partner
Your data. Inside your borders.
Regulated organizations require absolute data sovereignty. Soika delivers full AI capability deployed exclusively within your boundary — on-premise, air-gapped, or sovereign cloud. No external dependencies. No compromise.
National or regional cloud
Deployed into a national or regional cloud with in-country data residency, local operators and local support — sovereign AI in the strict sense.
- In-country residency
- Local operations
- Regulator-aligned
Private cloud & on-premise
Runs in your own data centre or VPC on hardware you own, integrated with your identity and network controls.
- Your hardware
- Your identity provider
- Your network policy
Air-gapped
Installed from signed offline media with no outbound connectivity, for classified and critical environments.
- No egress
- Offline updates
- Signed artefacts
Own the compute your AI runs on.
Tell us your capacity, location and timeline — we'll scope the cluster, the facility and the operating model with you.