What compute options does Aurora offer?

Aurora offers three compute deployment types. Virtual machines provide flexible vCPU and RAM sizing for general-purpose AI workloads, with full console, SSH, and API access. Managed Kubernetes clusters deliver GPU-aware pod scheduling, autoscaling, and full cluster lifecycle management. Dedicated instances provide single-tenant compute for workloads that require performance isolation, compliance separation, or dedicated GPU access. All three run on the same Aurora platform with the same management layer and 99.9% uptime SLA.

What GPU hardware does Aurora Compute support?

Aurora Compute supports NVIDIA H100, B200, and B300 GPU configurations, integrated with the Aurora GPU and AI product. GPU-aware Kubernetes scheduling places workloads on the correct hardware automatically without manual configuration. Dedicated GPU access is available on dedicated instances.

What is the difference between Aurora's three deployment models?

Aurora AI Platform deploys the Aurora software platform on the customer's existing hardware — rack space, owned GPU or CPU infrastructure. Managed Cloud deploys on Aurora-owned hardware, capital-light for the customer. Private AI IaaS covers net new infrastructure builds from the ground up, where Aurora procures, installs, and operates the full stack. All three use the same Aurora platform and the same 99.9% uptime SLA.

Can Aurora Compute be white-labeled and resold?

Yes. Aurora Compute is available as a fully white-labeled service for telcos, MSPs, and regional operators. Partners present compute services under their own brand — their logo, their domain, their pricing, and their customer relationships. Aurora operates the platform and remains invisible at every layer including the portal, domain, billing, and API endpoints.

How does Aurora handle Kubernetes management for compute workloads?

Aurora provides fully managed Kubernetes clusters with GPU-aware pod scheduling, horizontal and vertical autoscaling, and full cluster lifecycle management. Aurora handles the operations layer — cluster provisioning, GPU driver management, health monitoring, and incident response. Kubernetes API handoff is available for teams that want programmatic access to the cluster. Clusters integrate directly with Aurora storage and networking.

What security and encryption features are included with Aurora Compute?

Aurora Compute includes encryption in transit and at rest across all resources. BYO KMS is supported — customers can bring their own key management system such as HashiCorp Vault. IAM with role-based access control is built into every deployment. VPC and private networking isolate compute environments. Firewall and security groups provide granular traffic control at the instance and cluster level.

How quickly can Aurora Compute be deployed?

Aurora Compute deployments on existing infrastructure go live in 1 to 4 weeks depending on scope and configuration. New infrastructure builds take 8 to 16 weeks depending on the hardware configuration — CPU-only environments in approximately 8 weeks, 8-node GPU configurations in approximately 12 weeks, and 32-node GPU configurations in approximately 16 weeks.

How is Aurora Compute priced?

Aurora Compute is priced on a quote basis. Pricing depends on the deployment model, hardware configuration, scale, and operational requirements. Contact Aurora sales or book a technical demo to receive a scoped proposal — Aurora typically delivers indicative pricing within 1 to 2 calls.

Private Cloud Compute Infrastructure

Your Infrastructure. A Revenue Engine.

Operators and enterprises with rack space, owned hardware, or underutilized data center capacity have infrastructure but no platform to monetize it.

Aurora Compute solves that: virtual machines, Kubernetes orchestration, and dedicated instances — white-labeled, fully managed, ready to sell under your brand. Aurora operates the infrastructure. You own the customer relationship and the margin.

What Aurora handles:

Platform deployment and configuration
Day-to-day operations and monitoring
GPU drivers, Kubernetes, and autoscaling
SLA management and incident response
Billing infrastructure and API access

What you control:

Your brand — logo, domain, portal
Your pricing and packaging
Your customer relationships
Your revenue model — resell, subscription, or consumption
Your deployment region and data jurisdiction

Three Ways to Deploy Compute.

All three run on the Aurora platform — same management layer, same SLA, same operational model.

Virtual
Machines

Flexible vCPU and RAM sizing for general-purpose AI workloads. Spin up and down on demand. Full console, SSH, and API access.

Flexible vCPU + RAM configurations
Ubuntu 24.04 and other OS support
Console, stop, reboot, and refresh controls
Public and private IP addressing
API access and IAM built in

Kubernetes Clusters

Managed Kubernetes for containerized AI workloads. GPU-aware scheduling, autoscaling, and full cluster lifecycle management handled by Aurora.

Managed cluster provisioning and lifecycle
GPU-aware pod scheduling
Autoscaling — horizontal and vertical
Integrated with Aurora storage and networking
Kubernetes API handoff available

Dedicated Instances

Single-tenant compute for workloads requiring performance isolation, compliance separation, or dedicated GPU access. No noisy neighbour risk.

Single-tenant hardware allocation
Consistent, predictable performance
Suitable for regulated and sensitive workloads
Dedicated GPU access available
Custom sizing on request

What Comes With Every Compute Deployment.

Full Console Access

SSH connection, stop, reboot, refresh, and ticketing — all accessible via the Aurora portal, white-labeled under your brand.

Managed Kubernetes

Cluster provisioning, GPU scheduling, autoscaling, and full lifecycle management. Aurora handles the ops layer so your customers don't have to.

GPU-Aware Orchestration

Kubernetes scheduling that understands GPU topology — H100, B200, B300. Workloads land on the right hardware without manual configuration.

IAM & API Access

Configurable retention policies for regulatory requirements. Supports legal hold and audit-ready data governance.

Private Networking

VPC and private networking isolate compute environments. Firewall and security groups give granular traffic control at the instance and cluster level.

Integrated Storage

Compute instances connect directly to Aurora Storage — S3-compatible object storage, block volumes, and filesystem — without crossing a billing boundary.

White-Label Portal

Every compute resource your customers provision appears under your brand. Aurora is invisible at every layer — portal, domain, billing, and API endpoints.

99.9% Uptime SLA

Aurora operates what it deploys. Monitoring, incident response, and SLA management are part of every compute engagement.

Platform Specs.

Compute Types	Virtual Machines, Kubernetes Clusters, Dedicated Instances
vCPU	Flexible sizing — contact sales for configuration options
GPU Support	H100, B200, B300 — integrated with Aurora GPU & AI product
Operating Systems	Ubuntu 24.04; other OS on request
Networking	Public IP, private IP, VPC, firewall and security groups
Storage Integration	S3-compatible object storage, block volumes, filesystem, snapshots
Orchestration	Managed Kubernetes with GPU-aware scheduling and autoscaling
Access	Console, SSH, API, IAM
Encryption	In-transit and at-rest; BYO KMS supported
SLA	99.9% platform uptime (11 nines)
Deployment Models	Aurora AI Platform (your HW), Managed Cloud (Aurora HW), Private AI IaaS (new build)
Pricing	Quote-based — contact sales

What Operators Run on Aurora Compute.

Telcos & ISPs with Rack Space

Operators with existing data center infrastructure deploy Aurora Compute on their hardware. Aurora handles the platform — the operator white-labels and resells compute, storage, and AI services under their brand.

AI Inference
& Fine-Tuning

GPU-aware Kubernetes clusters for AI inference endpoints, model fine-tuning, and batch training workloads. Aurora handles orchestration — teams focus on models, not infrastructure.

Enterprises with Owned Hardware

Organizations with owned GPU or CPU infrastructure that want to run private AI workloads without routing through a hyperscaler. Aurora deploys the platform, manages operations, and keeps data in-region.

Regulated & Sovereign Deployments

Dedicated instances and private networking for organizations with data residency, compliance, or air-gap requirements. Compute stays in-region, under your control, with no hyperscaler routing.

Regional Cloud Providers

Operators building or expanding a regional cloud offering. Aurora provides the full compute stack — white-labeled, managed, and deployable in weeks rather than quarters.

GPUaaS
Providers

Organizations building a GPU-as-a-service offering. Aurora provides the infrastructure and platform — you set the pricing, own the customer, and earn on every GPU-hour.

Resell Aurora Compute Under Your Brand.

Aurora Compute is available as a fully white-labeled service for telcos, MSPs, and regional operators.
Your logo, your domain, your pricing. Aurora operates the infrastructure — you own the margin.

Learn more about the Aurora Partner Program

Let's Scope Your Deployment.

Whether you're deploying compute for your own AI workloads or building a service to resell, the technical demo is the fastest way to see exactly what Aurora can deliver for your infrastructure and your business model.

Book a Technical Demo

Walk through the platform with an Aurora engineer. We get to a proposal within 1-2 calls.

Connect with Aurora's Tech Team

Talk to Sales

PiB-scale, enterprise terms, or white-label deployment. Let's scope it.

Connect with Aurora

Compute Infrastructure Built for Private AI.