Core42 and Data Dynamics Partner to Enable Sovereign, AI-Ready Data Compliance in the UAE View announcement

Get In Touch

Accelerate Your AI Journey Today

Connect with us to unlock next gen GPU infrastructure for your AI and computational needs.
Exclude

Flexibility to Choose

Harness the power of NVIDIA, AMD, Qualcomm, Microsoft, or Cerebras by aligning each AI workload with the most suitable accelerator.

Exclude-1

Scale Without Borders

Scale instantly with thousands of GPUs globally, supporting everything from rapid prototyping to enterprise-grade training and inference.

Subtract

Built for Trust

Run mission-critical AI on secure, sovereign infrastructure designed to meet full regulatory requirements.

Who
is this for

From AI teams experimenting with the latest models to enterprises rolling out AI across thousands of users, Compass API Gateway provides the unified access, security, and control needed to scale with confidence.

Rectangle 3467305 (1)

Government/ Sovereign Cloud

Data sovereignty and compliance for public sector excellence across cloud, data, and AI.

Learn More
Screenshot 2026-04-17 at 4.07.06 PM 1

Large Enterprises

Scalable infrastructure for complex operations, multi-accelerator infrastructure and go from idea to production without friction.

Learn More
Rectangle 3467306

AI/ML Teams

Purpose-built for AI researchand production cloud with UAE-centric sovereign and security controls through Core42 Insight application

Learn More

Built for Choice, Scale, and Trust

Core42 AI Cloud brings together diverse accelerators, AI-optimized infrastructure, and production-ready inference in one platform at scale.

Multi-Accelerator Platform

Harness the power of NVIDIA, AMD, Qualcomm, Microsoft, or Cerebras by aligning each AI workload with the most suitable accelerator.

Global Scale, Sovereign Infrastructure

Access 86K+ GPUs across sovereign data centers globally. 

AI-Optimized Storage

Accelerate large-scale training and high-concurrency workloads with fast, resilient storage. 

Proven HPC Performance

Built on globally ranked Top500 and IO500 infrastructure for high-performance AI workloads. 

Core42 Compass Inference

Deploy leading models with low latency, built-in scalability, and sovereign-grade control. 

Frame 13 (1) Frame 18 Frame 14 Frame 16 Frame 17
ACCELERATORS

Peak Performance at Global Scale

Deploy AI workloads globally on NVIDIA, AMD, Cerebras, and Qualcomm accelerators with high-performance InfiniBand and Ethernet networking, orchestrated through Kubernetes or Slurm for peak efficiency.

GPUs-03
NVIDIA H100
Deploy Enquire
NVIDIA H100
Price: From $2.50/hr

The proven standard with InfiniBand networking for training and inference at scale.

Deploy Enquire
h200nvl-ari
NVIDIA H200
Enquiry
NVIDIA H200
Price: FROM $4.00/HR

Deliver breakthrough acceleration for large-scale AI training and long-context LLM inference workloads.

Enquiry
b200 (2)
NVIDIA B200
Enquire
NVIDIA B200
Price: FROM $5.00/HR

Next-generation Blackwell architecture with InfiniBand networking and NVLink interconnect for exceptional AI performance at scale.

Enquire
AMD Instinct MI300X OAM_IsoRight
AMD MI300X
Enquire
AMD MI300X
Price: FROM $3.50/HR

Power AI and HPC workloads with memory-optimized GPU, ROCm software stack and CDNA architecture optimized for inference acceleration and multi-chip efficiency.

Enquire
AMD MI355X-1
AMD Instinct MI355X
Enquire
AMD Instinct MI355X
Price: On Request

Runs large LLM models with a massive 288GB of HBM3E memory, reducing infrastructure footprint without sacrificing performance.

Enquire
Cerebras_Wafer_Scale_Engine_HD_1
Cerebras WSE-3
Enquire
Cerebras WSE-3
Price: On Request

Wafer-scale design removes interconnect bottlenecks, dramatically accelerating large-model training.

Enquire
CloudAI100_2
Qualcomm Cloud AI100 Ultra
Enquire
Qualcomm Cloud AI100 Ultra
Price: On Request

Optimized for performance per watt while delivering 870 TOPS of power in cost-efficient, sustainable high-volume inference.

Enquire
dgx-gb300-og 2
NVIDIA GB300 (Blackwell Ultra)
Enquire
NVIDIA GB300 (Blackwell Ultra)
Price: On Request

Built for next-generation liquid-cooled SuperPODs, powering real-time, multimodal AI at frontier scale.

Enquire
GPUs-03
NVIDIA H100
Deploy Enquire
h200nvl-ari
NVIDIA H200
Enquiry
b200 (2)
NVIDIA B200
Enquire
AMD Instinct MI300X OAM_IsoRight
AMD MI300X
Enquire
AMD MI355X-1
AMD Instinct MI355X
Enquire
Cerebras_Wafer_Scale_Engine_HD_1
Cerebras WSE-3
Enquire
CloudAI100_2
Qualcomm Cloud AI100 Ultra
Enquire
dgx-gb300-og 2
NVIDIA GB300 (Blackwell Ultra)
Enquire
Frame 19 (1) (1)

From Pilot to Production Inference

Core42 Compass unifies leading AI models, scalable inference, and built-in governance, eliminating the complexity of deploying AI at scale.

compass website 1 Group 2147205557 (1)
unified-model-access-icon

Unified Model Access

50+ industry leading models, including GPTs and open-source models across text, vision, speech, and embeddings all through one unified API.

Layer_1 (1)

Secure & Sovereign Deployment

Secure your GenAI deployments with in-country data residency, private endpoints, end-to-end encryption, guardrails, and enterprise-grade access controls.

Layer_1 (2)

Production-Scale Inference

High throughput processing capable of handling hundreds of millions of tokens in minutes from prototype to enterprise rollout.

Layer_1 (3)

Agentic Workflows

Build AI agents, multi-step orchestration, tool-calling systems, and autonomous workflows with production-ready frameworks.

Layer_1

Compass Playground

Experiment with prompts, test models, and validate performance before production.

redesigned-fine-tuning-icon

Fine-Tuning Services

Customize models to your domain with fine-tuning services and optional white-glove support from data preparation to deployment.

Group 2147205557 (1)
PLATFORM

One Platform. Built for Every AI Builder.

Train, fine-tune, and deploy agentic and inference workloads faster on a full-stack AI cloud with leading accelerators, integrated tools, and expert support.

AI-Optimized
Infrastructure
AI
Inference
AI Cloud Orchestration

GenAI Services

Accelerate innovation with services for agents, RAG, guardrails, and fine-tuning. Build and scale next-generation AI applications with confidence and speed.

GenAI

Core42 Compass Inference

Deploy and scale models in seconds with Core42 Compass, a unified platform for enterprise-grade inference. Access leading models through a single API, eliminate integration complexity, and run production workloads with low latency, sovereign control, and built-in reliability.

Model Hosting

AI OPS

Go beyond training with built-in AI lifecycle management. Customize and fine-tune models, monitor performance, and enforce governance with integrated AIOps to keep models reliable, compliant, and production-ready. 

AI Ops

Infrastructure as a Service

High-performance infrastructure with the freedom to choose from NVIDIA, AMD, Microsoft, Qualcomm, or Cerebras accelerators. Powered by diverse accelerators, AI-optimized storage, and high-speed InfiniBand and Ethernet networking, delivered as a fully managed platform for peak efficiency and performance. 

 

IaaS

Core
Benefits

Experiment, build, and scale GenAI seamlessly with Core42 Compass, combining startup speed with enterprise-grade control, 24×7 support, and 99.5% uptime.

Group 2147205553
Layer_1 (4)

Data Sovereignty

Keep AI models and data exchanges fully within UAE borders with comprehensive sovereign policies.

Subtract (1)

Powerful Unified API

Access leading AI models through a single API. No multi-vendor complexity, no performance trade-offs.

Subtract

Flexible Deployment Options

Choose cloud or on-premises deployment with infrastructure tailored to your unique security, performance, and compliance needs.

Group 2147205272

Future-Ready Architecture

Stay ahead of rapid AI advancements with a platform that evolves alongside your needs and maximizes long-term value.

Group 2147205558

AI Cloud by the numbers

Peak performance, proven at scale.

Deployments Managed
+150MW
GPU's Deployed/Planned
86 K
Worldwide
Global Data Centers
9
Worldwide
Exaflops
310
Worldwide
Top 500 HPC
#20
AMD MI300X globally
IO500
#3
Core42 Maximus-01 globally
Real-Time Responsiveness
20MS
Time To First Token
Production Ready
10B+
Tokens Processed Weekly

Start Fast. Scale on Your Terms.

Immediate, pay-as-you-go pricing via the AI Cloud console. Perfect for prototyping, model testing, and ML experiments that need to launch fast without commitment.

Get started with AI Cloud

How it works

Architecture overview

GenAI Services

Agent
RAG
Guardrails
Fine-tuning
Evaluation

Model Hosting & Inference

Inference
Model catalog
Model-as-a-service

AI Ops

Training
Model customization
Model governance

Infrastructure-as-a-service

Compute
nvidia
qualcomm
Cerebras_white-1
Microsoft_logo_(2012) (1) 1
AMD logo
Ultra-fast,
AI-optimized storage
High speed
networking
Throttling
User management
Billing
Metering
Managed
kubernetes & slurm
Access
management
Vector data
management

Trusted by Industry

 

Rectangle 20
image 4
image 5
image 6
image 7
img-2aa3f2704899bd9b
image 9
Rectangle 20
image 4
image 5
image 6
image 7
img-2aa3f2704899bd9b
image 9
image 11
image 12
image 13
image 14
image 15
image 16
image 17
image 11
image 12
image 13
image 14
image 15
image 16
image 17
image 18
image 19
image 20
image 21
image 22
image 23
image 24
image 18
image 19
image 20
image 21
image 22
image 23
image 24

Resources

Start your AI journey with the insights, tools, and resources to turn ideas into production-ready solutions.

2 2

AI Cloud Overview

Download to learn about AI Cloud.

1 3

AI Cloud Platform Demo

Watch the platform demo to learn more about AI Cloud capabilities.

Compass_resources (1) (Medium)

Compass
brochure

Download Compass brochure to learn how to simplify your AI journey

Compass_platform_2

Compass Platform
Demo

See Core42 Compass in action with a guided platform walkthrough.

FAQs about Core42 AI Cloud

What is Core42 AI Cloud and how is it different from a standard GPU cloud?

Core42 AI Cloud is a full-stack, AI-native cloud platform, not a general-purpose cloud with GPUs added. It integrates heterogeneous compute (NVIDIA, AMD, Qualcomm, Cerebras), AI-optimized storage, high-speed networking, unified orchestration (bare metal, Kubernetes, SLURM), and Core42 Compass inference into one platform. The result is an environment built for the full AI lifecycle: training, fine-tuning, inference, deployment, and continuous refinement without operational handoff friction between stages.

What accelerators are available, and can I mix different GPU types?

Core42 AI Cloud supports NVIDIA, AMD, Qualcomm, Cerebras, and Microsoft. Mixed accelerator fleets are supported - you can align each workload with the most suitable hardware without disrupting your broader infrastructure strategy. This is a deliberate architectural choice: frontier AI is not monolithic, and different workloads demand different memory architectures and scaling behaviors. You are not locked into any single vendor's roadmap.

What consumption models and pricing options are available on Core42 AI Cloud?

Core42 offers flexible consumption models including on-demand GPUaaS, large scale clusters, and inference-based pricing through Core42 Compass such as pay-as-you-go or tokens-per-minute with full cost transparency. 

How does Core42 AI Cloud support large-scale AI training and high-performance workloads?

The platform supports everything from on-demand GPU instances to large-scale clusters, with high-speed networking, AI-optimized storage, and managed orchestration through Kubernetes and Slurm, enabling efficient training and high-concurrency workloads at scale.

How does Core42 ensure data sovereignty, security, and compliance?

Core42 AI Cloud operates across sovereign data centers, with in-country data residency, encryption at rest and in transit, and enterprise-grade governance controls, ensuring compliance for regulated industries and national-scale AI initiatives.

Which model providers are available on Compass?

We regularly onboard new closed-source and open-source releases and currently offers models from 12 providers including OpenAI, Anthropic, Cohere, Meta, Mistral, Stability AI, xAI, DeepSeek, Qwen, MBZUAI, Liquid AI, and Inception. The model provider is transparently disclosed in the platform, and customers explicitly choose which model to use.

Does Compass store or use customer data for training?

No. Customer data is processed transiently for inference only. Compass does not store prompts or outputs, does not reuse data, and does not train or fine-tune models using customer inputs. All data remains fully owned by the customer.

How does Compass support governance and access control?

Compass provides platform-level governance through API key management, role-based access control, audit logs, usage monitoring, and billing transparency. Customers can monitor model usage, manage users at admin or department level, and track activity through the Compass portal or APIs.

How easy is it to integrate Compass?

Compass offers a single unified API that allows developers to integrate AI models directly into applications. This streamlines development, reduces integration complexity, and enables rapid deployment across legacy and modern systems.

Compass API protocol is compatible with OpenAI and Azure OpenAI.

What service levels and support are available?

Compass provides uptime-focused SLAs with 99.5% availability. Customer support is available 24×7.

How does Compass improve time and cost efficiency?

Enterprises can accelerate innovation by using pre-built, pre-trained models without investing in complex infrastructure or specialized machine learning expertise. The as-a-service model reduces upfront costs while enabling scalable, production-grade AI deployment.

Ready to Accelerate
Your AI Journey?

Deploy AI workloads globally across sovereign data centers with built-in scale, security, and accelerator choice.