Public Cloud/Compute/GPU Cloud Instances
GPU Compute

GPU Cloud Instances

High-performance GPU instances powered by NVIDIA A100 and H100. Purpose-built for AI/ML training, deep learning, rendering, and scientific computing.

Instance Types & Pricing

Monthly pricing shown. Hourly billing available at ~1.5x monthly rate.

NVIDIA A100 40GB

Ideal for AI training and inference workloads

InstancevCPURAMStorageBandwidthPrice/mo
gpu.a100.2x24240 GB1 TB NVMe-₹150/hr
gpu.a100.4x48480 GB2 TB NVMe-₹295/hr
gpu.a100.8x96960 GB4 TB NVMe-₹580/hr
gpu.a100.1x12120 GB500 GB NVMe-₹755/hr

NVIDIA A100 80GB

High Memory

Extended memory for large model training

InstancevCPURAMStorageBandwidthPrice/mo
gpu.a100-80.1x12120 GB500 GB NVMe-₹95/hr
gpu.a100-80.2x24240 GB1 TB NVMe-₹185/hr
gpu.a100-80.4x48480 GB2 TB NVMe-₹365/hr
gpu.a100-80.8x96960 GB4 TB NVMe-₹720/hr

NVIDIA H100

Latest

Latest generation for cutting-edge AI workloads

InstancevCPURAMStorageBandwidthPrice/mo
gpu.h100.1x16200 GB1 TB NVMe-₹165/hr
gpu.h100.2x32400 GB2 TB NVMe-₹325/hr
gpu.h100.4x64800 GB4 TB NVMe-₹640/hr
gpu.h100.8x1281.6 TB8 TB NVMe-₹1,250/hr

NVIDIA T4

Budget

Cost-effective for inference and light training

InstancevCPURAMStorageBandwidthPrice/mo
gpu.t4.1x416 GB200 GB NVMe-₹25/hr
gpu.t4.2x832 GB400 GB NVMe-₹48/hr
gpu.t4.4x1664 GB800 GB NVMe-₹92/hr

Features

NVLink Interconnect

High-bandwidth GPU-to-GPU communication for distributed training.

Pre-installed CUDA

Ready-to-use with CUDA, cuDNN, and popular ML frameworks.

Jupyter Integration

One-click Jupyter notebook access for interactive development.

Spot GPU Instances

Save up to 70% with interruptible GPU instances for training.

Common Use Cases

LLM Training

Train large language models with multi-GPU support

Computer Vision

Image classification, object detection, segmentation

3D Rendering

Real-time rendering and ray tracing workloads

Scientific Computing

Molecular dynamics, weather simulation, HPC

Video Processing

Transcoding, AI upscaling, real-time processing

Inference at Scale

Deploy ML models for production inference

Start Training Today

Get GPU access in minutes. No long-term commitment required.