GPU capacity for AI teams
Reserve GPU instances by the hour for fine-tuning, inference serving or rendering. Attach persistent NVMe volumes so checkpoints survive between jobs.
For steady-state workloads, monthly reservations carry a lower effective rate than on-demand pricing.
Data residency for AI workloads
Training data and model checkpoints remain inside Malaysia. This matters when your dataset contains personal data regulated under PDPA, or when a client contract requires in-country processing.
- Hourly and monthly reservation
- Persistent NVMe volumes for checkpoints
- Private networking between GPU nodes
- Local support during MYT business hours