MashupAbility//OpenSource

GPU CLOUD INFRASTRUCTURE · PROVIDER DIRECTORY

Where open models run.

A source-led map of the hyperscalers, neoclouds, specialist GPU providers, and sovereign AI clouds that can carry open and open-weight model workloads.

85

provider entries

4

infrastructure lanes

62

specialist providers

15

sovereign providers

INFRASTRUCTURE LANE

Traditional hyperscalers

4 entries

Full-stack cloud providers offering general-purpose infrastructure alongside GPU instances and managed AI services.

Microsoft Azure

Global

GPU VMs + managed AI platform

Official ↗

Oracle Cloud Infrastructure

Oracle / OCI

Global

Bare metal + GPU VMs

Official ↗

Amazon Web Services

AWS

Global

Accelerated EC2 + managed AI

Official ↗

INFRASTRUCTURE LANE

Neocloud giants

4 entries

Large, heavily capitalized AI-infrastructure specialists operating substantial accelerator fleets and dedicated clusters.

CoreWeave

North America + Europe

AI-native cloud + clusters

Official ↗

Lambda

Formerly Lambda Labs

North America + Europe

On-demand GPUs + clusters

Official ↗

Crusoe

North America + Europe

AI cloud + dedicated infrastructure

Official ↗

Nebius

United States + Europe

AI cloud + managed services

Official ↗

INFRASTRUCTURE LANE

Specialist and emerging GPU clouds

62 entries

Regional, bare-metal, PaaS, marketplace, and boutique providers spanning training, fine-tuning, and inference workloads.

Together AI

North America + multi-region

Inference platform + GPU cloud

Official ↗

Applied Digital

United States

AI/HPC campuses + contract capacity

Official ↗

Sustainable Metal Cloud

SMC / Firmus

Singapore + APAC

Sustainable GPU cloud

Official ↗

TensorWave

United States

AMD AI cloud + dedicated clusters

Official ↗

IREN

Formerly Iris Energy

North America

Contract GPU cloud + AI data centers

Official ↗

Denvr Dataworks

Denvr Data

Canada + North America

GPU cloud + clusters

Official ↗

Hut 8

North America

AI data centers + contract infrastructure

Official ↗

Lamini

North America

Enterprise LLM platform + GPU cloud

Official ↗

Genesis Cloud

Europe

GPU instances + dedicated clusters

Official ↗

Akamai Cloud

Linode / Akamai

Global

GPU instances + general cloud

Official ↗

Horizon Compute

United States; Japan expansion

On-demand + reserved GPU clusters

Official ↗

HPE GreenLake

Global + customer locations

Private cloud as a service

Official ↗

Latitude.sh

Americas + Europe

Bare metal GPU cloud

Official ↗

Gcore

Global + Europe

GPU cloud + inference platform

Official ↗

Nscale

Europe + North America

AI cloud + dedicated clusters

Official ↗

Valdi

Global marketplace

Multi-provider GPU marketplace

Official ↗

Hyperstack

NexGen Cloud; duplicate listing merged

Europe + North America

On-demand GPU cloud

Official ↗

Lepton AI

Global marketplace

AI platform + federated GPU capacity

Official ↗

Green AI Cloud

Sweden + Europe

Sustainable GPU cloud

Official ↗

TensorDock

Global marketplace

On-demand GPU marketplace

Official ↗

Turboscale

Global clusters

GPU cloud + custom infrastructure

Official ↗

Cirrascale

North America

Managed AI cloud + dedicated systems

Official ↗

Fluidstack

North America + Europe

Contract AI cloud + clusters

Official ↗

Prime Intellect

Global marketplace

Distributed compute marketplace

Official ↗

Taiga Cloud

Northern Data

Europe

GPU cloud + clusters

Official ↗

Bitdeer AI Cloud

North America + Asia

AI cloud + data-center capacity

Official ↗

Ubitus

Asia-Pacific

GPU cloud + managed AI infrastructure

Official ↗

Hypertec Cloud

North America + Europe

Private GPU cloud + managed infrastructure

Official ↗

EscherCloud

EscherCloudAI

Netherlands + Europe

GPU cloud + AI infrastructure

Official ↗

Arc Compute

Global deployments

Designed + managed GPU infrastructure

Official ↗

FlexAI

Duplicate listing merged

Europe + North America

Workload platform + multi-cloud compute

Official ↗

Backprop

Backprop Co

North America

GPU cloud + dedicated capacity

Official ↗

JarvisLabs.ai

Jarvis Labs AI

North America

On-demand GPU instances

Official ↗

Sify Technologies

India

Enterprise cloud + GPU infrastructure

Official ↗

GMI Cloud

North America + Asia

GPU cloud + inference platform

Official ↗

Dihuni

North America

GPU cloud + systems integration

Official ↗

Gompute / Advania

Formerly atNorth Gompute

Nordics + Europe

Private HPC cloud + bare metal

Official ↗

Patmos

United States

AI data centers + cloud + colocation

Official ↗

Atos

Global + Europe

Managed cloud + HPC infrastructure

Official ↗

Atlas Cloud

United Kingdom + Europe

GPU cloud + dedicated clusters

Official ↗

OVHcloud

Global + Europe

Public-cloud GPU instances

Official ↗

FastGPU

Needs verification

Supplied name resolves to scheduling software, not a verified cloud

Review source ↗

CR8DL

United States

GPU cloud + virtual data center

Official ↗

INFRASTRUCTURE LANE

Sovereign AI clouds

15 entries

Localized infrastructure designed around regional control, data-residency requirements, national AI programs, or domestic-language ecosystems.

Yotta Shakti Cloud

Yotta

India

Sovereign AI cloud + GPU infrastructure

Official ↗

YTL AI Cloud

YTL

Malaysia + Southeast Asia

Sovereign AI cloud + supercomputing

Official ↗

FPT AI Factory

FPT

Vietnam + Japan

GPU cloud + bare metal + inference

Official ↗

Core42

G42 company

UAE + sovereign global regions

Sovereign AI cloud + multi-accelerator compute

Official ↗

Reliance Jio

Reliance

India

AI cloud infrastructure; availability evolving

Official ↗

Tata Communications Vayu

Tata

India

Sovereign GPU cloud + managed AI

Official ↗

Singtel RE:AI

Singtel

Singapore + Southeast Asia

Sovereign AI cloud + GPUaaS

Official ↗

Krutrim Cloud

Ola Krutrim

India

AI-native cloud + GPU VMs + bare metal

Official ↗

GreenNode

Vietnam + Thailand

Sovereign AI cloud + GPU instances

Official ↗

KDDI GPU Cloud

KDDI

Japan

Secure GPUaaS + domestic networking

Official ↗

RUTILEA

Japan

GPU data centers + integrated AI services

Official ↗

GPUSOROBAN

Highreso

Japan

Domestic on-demand GPU cloud

Official ↗

SAKURA Koukaryoku

SAKURA internet

Japan

Bare metal + container + VM GPU cloud

Official ↗

Omniva

United States; Middle East + Europe plans

AI infrastructure startup; access model not public

Official ↗

SCOPE NOTES

A map, not a ranking.

Categories describe market position and operating model; they do not score price, reliability, security, sovereignty, or live accelerator availability. Verify those factors directly before choosing a provider.

The supplied emerging list contained 64 rows. Hyperstack and FlexAI each appeared twice, so this directory reports 62 unique specialist entries—not 78. The sovereign list contains 15 entries, not 14. FastGPU remains visible but is marked for review because the supplied name resolves to GPU scheduling software rather than a verified cloud.