Skip to main content
This guide takes a datacenter operator from nothing to a running, branded, multi-tenant GPU cloud — including where every byte of your data lives.

​
The path

1

Choose a deployment model

Managed, sovereign, or on-prem. This decides where your data lives and who operates the stack. → Deployment models
2

Create your operator account

An invite code that names your datacenter makes you its first admin. → Create your account
3

Add your team

More operator seats, and what they can each do. → Your team
4

Issue an enrollment token

Named, revocable, revealed once. → Enrollment tokens
5

Install the agent

One command, the same one fleet-wide. Nodes appear within two minutes. → Install the agent
6

Create tenants and invite customers

One tenant per customer company. → Tenants
7

Run it

Reading a node, reassigning, offline handling, drills. → Day-2 operations

​
What you need before you start

An operator invite code

Issued by NeoChamber during onboarding. It names your datacenter.

GPU servers

Ubuntu 22.04 or 24.04+, NVIDIA driver installed, outbound HTTPS on 443.
That is the entire prerequisite list. No inbound network access, no Docker, no runtime on the machines, and no infrastructure of your own unless you chose a sovereign or on-prem deployment.

​
Commercial terms

Pure revenue share: 5% of GPU services revenue. No upfront licence, no per-seat fees, no integration charges. Idle-compute resale income is counted under the same rate. NeoChamber deploys, hosts, and operates the full stack in the managed model, and earns only when you earn.