A100 SXM4 with 80GB HBM2e — the workhorse for training, fine-tuning, and inference at a lower price point than H100. INR-native billing, sub-2-minute provisioning, GST invoices.
99.9% SLA Uptime
Backed by contract. Auto credits if we miss it.
Per-Minute Billing
Pause anytime. Billing stops instantly. Resume same state.
< 2 Min Provisioning
GPU live and SSH-ready in under 2 minutes. No queues.
From instant Jupyter workbenches to full root-access VMs. Switch anytime.
A full breakdown of capabilities across every compute environment.
The A100 (Ampere architecture) handles most training, fine-tuning, and inference workloads well, with 80GB of HBM2e memory. It's the go-to choice when you don't need H100-level throughput but still want serious multi-GPU capability.