Yotta Data Services logo

GPU Cloud Provider

Yotta Data Services

GPUs
5
Countries
1

GPU Marketplace

NVIDIA T4On-Demand
$0.33/hour
NVIDIA A40On-Demand
$2.11/hour
NVIDIA L4On-Demand
$2.20/hour
NVIDIA L40SOn-Demand
$2.25/hour
NVIDIA L40SOn-Demand
$2.30/hour

Estimates onlyrates are collected automatically from public provider pages and may be out of date. Prices vary by region, commitment term, and availability, and typically exclude storage, egress, and tax. Confirm current pricing with the provider before purchasing. Last collected .

Company Profile

Company TypeIndia’s first and truly indigenous AI-HPC Cloud
Provider TypeAI-HPC Cloud

Infrastructure

GPU FleetThousands of NVIDIA H100 or L40S GPUs
Bare MetalShakti Cloud Bare Metal GPU instances provide access to dedicated GPUs
organisations, businessesAI researchers, and numerous startups

Compute & Deployment

On-DemandOn-demand infrastructure, customisable and secure
Bare MetalDedicated Physical GPUs
Container-Basedruns natively on Kubernetes and Slurm
Kubernetesruns natively on Kubernetes and Slurm

GPU Hardware

Latest GenNVIDIA H100 Tensor Core GPU, NVIDIA L40S GPU
InfiniBandInfiniBand Quantum-2 400 Gbps networking (with 3200 Gbps through InfiniBand Quantum-2 switches)

Pricing Model

Per Hour$4.76/hr
SubscriptionInfrastructure As A Service; Platform As A Service; Software As A Service
Pay-as-you-goShakti Cloud offer pay-as-you-go pricing models, allowing users to pay only for the resources they consume without any upfront costs.

Performance & Scaling

Multi-Node TrainingTrain Large-Scale Models Across Thousands of NVIDIA H100 or L40S
Max Cluster SizeTrain Large-Scale Models Across Thousands of NVIDIA H100 or L40S
Elastic ScalingShakti Cloud Bare Metal GPU instances are scalable, allowing users to add or remove instances as needed to handle varying workloads.
InfiniBandLeverage massive GPU clusters based on NVIDIA reference architecture, featuring InfiniBand Quantum-2 400 Gbps networking for exceptional node-to-node bandwidth. Achieve 3200 Gbps connectivity through InfiniBand Quantum-2 switches.
Perf IsolationShakti Cloud Bare Metal instances provide a higher level of isolation and security, as users have exclusive access to the physical hardware, reducing the risk of security vulnerabilities associated with virtualisation.
Noisy NeighborShakti Cloud Bare Metal instances provide a higher level of isolation and security, as users have exclusive access to the physical hardware, reducing the risk of security vulnerabilities associated with virtualisation.

Developer Experience

Data Center Locations

Coverage

CountriesIndia
the IndianAsian, and global markets

Compliance Regions

India RegionIndia
Datacenter Locations

Key Strengths

Indigenous AI-HPC cloud in collaboration with NVIDIA
very large scale (16 Exaflops claimed)
large fleet of NVIDIA H100 and L40S GPUs
high-speed InfiniBand Quantum-2 networking (400 Gbps per link
3200 Gbps fabric)
NVME and SSD storage
bare-metal dedicated GPUs with low-level access
runs on Kubernetes and Slurm
APIs and management tools
built-in redundancy and failover for high availability.

Known Limitations

No spot instance pricing mentioned
no reserved-instance options mentioned
no free tier mentioned
no region/multi-region or multi-cloud details
no numeric SLA percentage provided
no PUE or energy/green commitments provided
no SDK language list, no Jupyter support mentioned
no named customers or funding/investors listed.

Additional Information

Core Proposition

India’s Largest Super Computer of 16 Exaflops AI computing power with NVIDIA H100 Tensor Core GPUs

Last updated August 2026. Information subject to change.