Dedicated by design.
One customer. A dedicated cluster.
A delivery plan built around both.
AI teams need to know what capacity they can rely on, what it will cost, and who is responsible for delivering it. For workloads that require a dedicated cluster, that commitment extends well beyond the hardware.
Keraunos Compute Inc. is a US company developing dedicated GPU infrastructure for South America, starting in São Paulo. The model brings single-tenant clusters into established colocation facilities, with customer commitments guiding each deployment.
The focus is specific: bring together the equipment, the facility, the financing and the operating team to deliver the capacity a customer has contracted for.
01Dedicated capacity.
A dedicated cluster gives one customer a defined resource to plan around. Its configuration, service scope and availability commitments should follow that customer’s workload.
That is the product Keraunos is building: GPU capacity reserved for a single customer under a long-term agreement. The relationship begins with understanding what that customer needs to run.
02Established facilities.
The initial deployments are planned for existing colocation data centres in São Paulo. This puts established facilities at the foundation of the service and allows Keraunos to concentrate on the compute infrastructure and its delivery.
Specialist providers are expected to support installation and operations under Keraunos oversight. The company retains responsibility for customer delivery and service performance.
03Customer commitments first.
The model begins with a customer contract. Hardware procurement follows a signed agreement, the required prepayment, confirmed project funding and site readiness.
Each cluster is intended to be financed against its own customer contract. Planning the commercial, financing and delivery arrangements together keeps the deployment tied to a specific customer and a defined scope.
04Scale through execution.
The first priority is to deliver the initial cluster and operate it well. That requires the equipment, the site, the operating team and customer acceptance to come together.
Expansion should follow a process the company can repeat: contract, procure, commission and operate. The quality of those deployments will determine the pace of growth.
The ambition is a dedicated compute platform for South America, built one customer and one cluster at a time.
That is the work ahead of Keraunos.
Keraunos Compute Inc.
DEDICATED GPU INFRASTRUCTURE / SOUTH AMERICA
Start a conversation