Private clusters
Your own hypervisors. Our operations team.
A private cluster is a set of dedicated hypervisors reserved for your organisation alone. You get the isolation of bare metal and keep the console, CLI and API you'd use for any Edge VM — we look after the hardware.
Most clusters are ready within 48 hours.
How it works
Dedicated hardware without a hardware project
You never rack a server or sign a colocation contract. We build the cluster; you deploy to it.
- 01
We provision your cluster
Dedicated hypervisors go live in the location you choose. IP pools are assigned and the cluster is configured for your account. - 02
You deploy through Edge
Create VMs from the console, CLI or API exactly as you would on shared compute. Your cluster is selected automatically. - 03
We run the hardware
Maintenance, network uplinks, power and cooling are ours to handle. Your team works on the workloads.
What you get
Everything in shared compute, plus the walls
Dedicated hardware
Hypervisors that run only your VMs — for compliance, predictable performance and security reviews.Complete isolation
Your cluster is invisible to other customers and your VMs land only on your hardware. No contention.Same interface
The Edge console, CLI and REST API you already know. No separate tooling to learn.CLI and API automation
Deploy, scale and manage programmatically, from CI or your own tooling.Agentic API
Self-describing endpoints with natural-language summaries, so AI agents can operate the cluster safely.Flexible IP pools
Dedicate IP pools to your cluster or share them across clusters to fit your network design.
No new tooling
Nothing in your pipeline changes
Private clusters use the same API as shared compute. Existing scripts, CI jobs and agents keep working; only the placement changes underneath them.
- Console, Edge CLI, REST API and agent API
- Same VM features: backups, private networks, firewall rules
- Per-organisation cluster selection — no flags to forget
deploy.sh
# Exactly the same call as shared compute.
# Your organisation's cluster is selected automatically.
curl -X POST https://edge.network/api/compute/vms \
-H "Authorization: Bearer ek_live_..." \
-H "Content-Type: application/json" \
-d '{
"name": "ledger-api-01",
"region": "london",
"osTemplate": "ubuntu-24.04",
"vcpu": 16,
"ram": 64,
"disk": 500
}'Built for
When shared isn't an option
Regulated industries
Financial services, healthcare and government workloads that require dedicated infrastructure.High-performance computing
Consistent throughput with no neighbours competing for the same hardware.Data sovereignty
Keep data on hardware in a specific jurisdiction, with full control over where VMs run.Enterprise SaaS
Run your platform on guaranteed, isolated resources your customers can audit.
Shared or private
Which one do you need?
Most teams start on shared compute. Move to a private cluster when isolation becomes a requirement, not a preference.
| Private cluster | Shared compute | |
|---|---|---|
| Hardware | Dedicated hypervisors | Shared hypervisors |
| vCPU and RAM | Dedicated | Dedicated |
| Hardware reserved for you alone | ||
| Location | Your chosen location | Any of 33 regions |
| Console, CLI and API | ||
| Hardware operations | Managed by Edge | Managed by Edge |
| Time to start | Most within 48 hours | Under 60 seconds |
| Pricing | Quoted for your cluster | Per resource, hourly |
FAQ
Private cluster questions
How long does a private cluster take to set up?
Most clusters are ready within 48 hours of agreeing the requirements.
Do I need different tools to manage it?
No. You use the same Edge console, CLI, REST API and agentic API. VMs you create are placed on your cluster automatically.
Who maintains the hardware?
Edge does — hardware maintenance, network uplinks, power and cooling. You manage the VMs and what runs on them.
How is a private cluster priced?
It depends on the size and location of the cluster. Tell us what you need and we'll send a proposal.
Tell us what the cluster needs to do
Share the location, capacity and compliance requirements. We'll configure the cluster — most are ready within 48 hours.
No commitment to talk. Shared compute is available right now.