Compute

AI Compute

Dedicated GPU nodes on infrastructure we built and still run ourselves.

You take the capacity; the racks, the fabric, the drivers and the night shift stay ours. Every cluster we offer capacity on was sized, racked, cabled and commissioned by our own engineers, and the same teams are the ones watching it now. Day 2 operations are running today in Indonesia, Singapore and Malaysia.

A lit aisle of GPU compute racks in a white data hall, cabinet doors closed and status lights running the length of the row.
Scope

What the engagement covers

  • Whole GPU nodes reserved to you, not shared with another tenant

  • InfiniBand or RoCEv2 / RDMA interconnect between nodes

  • Enterprise storage, sized for training workloads

  • Workloads pinned to a named facility in a named market

What you can run on

Platforms in service — configuration quoted per engagement

LayerIn service, and how to get it
GPU platformMultiple GPU platform generations in service. Available capacity on request.
InterconnectInfiniBand, or RoCEv2 / RDMA over Ethernet. Fabric per cluster on request.
StorageEnterprise storage, sized for model training workloads. Capacity on request.
Node configurationSized against the workload. On request.
RegionSingapore, Malaysia, Indonesia and Taiwan. Available capacity on request.
Commercial termsTerm, minimum commitment and rate on request.

We do not publish a rate card. Capacity, node configuration and term are quoted against the workload — tell us what you need to run, and which market it has to sit in.

Get started

Tell us what you need to run.

Share your workload requirements, market focus, and timeline. Our expert teams across Taiwan, Singapore, Malaysia, and Indonesia will design the optimal compute service—leveraging our existing GPU capacity or building a dedicated infrastructure tailored to your needs.

Talk to our team →