AI Infrastructure
Compute & GPU access
The Centre is not simply a GPU centre. Compute is an enabling resource made available to selected AI startups, allocated deliberately rather than handed out by default.
Available categories
- GPU compute
- Cloud compute
- CPU resources
- Development environments
- Storage and data infrastructure
- Technology partner credits and resources, where applicable
Allocation principles
Needs-basedTechnically justifiedProgramme-linkedSubject to availabilityPeriodically reviewedUtilisation monitored
What you'll be asked to indicate
To request compute, startups indicate their actual requirement rather than a blanket ask:
Training workloadInference workloadPreferred GPU classGPU hoursStorageCPU / RAMCloud requirementsDevelopment stackDataset sizeExpected utilisation
Subsidised pricing
Subsidised compute access: rates to be announced, based on approved allocation policy.
Compute access is subsidised, not free. This page will be updated with the approved rate card once it is finalised.