We install it, cluster it so the estate keeps running when a node dies, and put Proxmox Backup Server behind it. Then we restore one of your machines and boot it while you watch, because a backup job that reports success has proved nothing. On your hardware or ours.
Virtualisation licensing is sold per socket, per core or per host, and the number changes when the vendor decides it should. Proxmox removes that conversation.
What replaces it is your own responsibility for the thing underneath everything else you run. That is a fair trade if the cluster survives losing a node and the backups have been restored at least once.
Most have not been. A backup job that reports success every night proves that a job ran, and nothing else.
Every step is a fixed price with its own deadline. You can stop after any of them and owe nothing further.
A virtual machine or a container built to a standard shape, with its resources, network, storage and access set the same way every time.
A new Linux VM on your own Proxmox, handed over finished: your keys in, root and password logins shut, and only the ports you named answering — each of those proved by trying it from a machine you choose rather than by reading a config file.
8 weekday hours
A new Linux container on your own Proxmox, handed over finished and unprivileged by default: your keys in, root and password logins shut, only the ports you named answering — and the isolation you actually got written down instead of assumed.
8 weekday hours
Nodes joined into a cluster with a quorum that holds, shared storage, and live migration proved by moving a running machine between hosts in front of you.
Proxmox Backup Server installed, retention set, and a whole guest restored from the archive while you watch. The restore is the deliverable, not the schedule.
The management interface restricted, its own protections switched on, and a way back in proved before we hand over so a hardening pass cannot lock you out of your own estate.
For an estate somebody else built: what is running, what is exposed, what is not backed up, and what will break next, written down with severities.
A cloud-init template so a new machine arrives configured rather than built by hand, which is what makes the fiftieth one as cheap as the second.
The steps above that are not marked optional come to US$636 before any per-person or per-machine work, which is priced per unit and shown on each task page.
Patching, expiring certificates, disks creeping toward full and a restore test nobody has run: none of it is urgent until the day it takes something down.
You can keep that work in-house and buy days when you need them, or hand the whole estate over. We can also run the cluster on our own hardware, in the United States or the European Union, with the export that takes it back sitting in your hands from the first month.
A senior system administrator works a full day on your server backlog, from the patch level nobody has checked to the backup nobody has restore-tested. Book one day or a run of dates, and each day ends with a written account of every hour.
Up to 10 Linux VMs patched every month in a weekday window you choose. Each one is snapshotted first, rebooted and checked, and two urgent runs a month cover the fixes that cannot wait.
Your Proxmox desktop estate run for you every month, down to one of your own machines restored from backup, booted and timed. The report shows you the evidence.
Prices stay fixed because the edges are drawn before you book, not after.
Describe your situation and we will tell you which of these steps you need and which you can skip.