Reserve or run on-demand — bare-metal performance with FluidCore's sovereignty guarantees. All clusters include NVLink fabric, InfiniBand networking, and dedicated NVMe storage.
| Configuration | GPU Model | On-Demand HR | 1-Month Reserved | 3-Month Reserved |
|---|---|---|---|---|
| 4× A100 80GB node | A100 SXM | ₹960 | ₹820 | ₹720 |
| 8× A100 80GB node | A100 SXM | ₹1,850 | ₹1,480 | ₹1,295 |
| 8× H100 80GB node | H100 SXM5 | ₹4,200 | ₹3,360 | ₹2,940 |
| Custom cluster (16+ GPUs) | A100 SXM | Any | Custom | Dedicated SLA |
Data sovereignty, At-rest encryption, Real-time usage API, S3-compatible SDK are included with all Storage Plans
400 Gbps interconnect for multi-GPU jobs
No virtualisation overhead, India-resident hardware
Docker, Kubernetes, SLURM pre-configured images
High-speed NVMe attached, survives instance restarts
Up to 70% off on preemptible workloads
Utilisation, memory, PCIe bandwidth in real time