Your own server room is a large CAPEX
Leadership sees it as a capital outlay before anyone powers on the first virtual machine. Hosts, arrays, racks, UPS, precision cooling, physical access control and connectivity are capital you freeze, depreciate and then service for years. Add floor space, power, on-call cover and the risk that in a few years the kit is too small — or too weak — for the next wave of workloads.
On the software side Broadcom moved VMware to subscription. Even if you keep the hypervisor on-prem, licences do not return to a buy-once model. The investment budget swells with hardware; the operating budget swells with support. It is a poor moment to add another server room inside your own office.
Public cloud is often the right answer for bursty traffic. It is not automatically cheaper for steady, local workloads — networking, database licences and data egress can eat the saving the board was looking for.
What changes in an OPEX model
We host your virtual machines on a Proxmox VE cluster in a professional Warsaw data centre. You do not buy hosts as fixed assets. You pay on a cycle for capacity, power, cooling, connectivity and care — like any other operating service. Hardware, the rack and physical security sit with the facility and our contract, not with your CAPEX.
Proxmox VE is open-source software (AGPL). An optional support subscription from Proxmox GmbH is a fraction of VMware list price and is not a condition for the cluster to run. The hypervisor, high availability and Proxmox Backup Server can be billed together with rack space, instead of as a separate capital project.
- No server purchase to start — you do not wait on a hardware tender before production can move.
- Capacity without buying a rack — add vCPU, memory and disk as the business grows; you do not buy hardware “just in case”.
- One operating invoice — facility, power, connectivity and platform in one cost that finance reads as OPEX.
- Leave VMware in the same move — we can combine virtual-machine migration with the move to Warsaw; the path is in the VMware to Proxmox guide.
How finance sees it
CAPEX is capital expenditure: servers, racks, cooling, cabling. OPEX is the running cost: the service, power at the operator, the subscription. Your own server room mixes both — a large outlay on day one and standing bills afterwards. Hosting a Proxmox cluster in a data centre shifts the weight onto OPEX: the company does not buy hosts onto the asset register; it pays for a service. Whether a given contract is operating spend, a lease or something else is a classification for your finance team.
| Line | Own server room (CAPEX) | Proxmox in Warsaw (OPEX) |
|---|---|---|
| Hardware | Buy hosts, disks, network — depreciate for years | Capacity in the service, without buying servers as assets |
| Building | Room, racks, cooling, UPS, physical security | Professional data centre — those layers sit in the rack fee |
| Hypervisor | VMware subscription or another licensing project | Proxmox VE: open source, optional vendor support |
| Power and cooling | Your energy bill plus cooling maintenance | On the facility side, inside the service charge |
| People | Rack on-call, travel, spare parts | Platform care and incident procedure in the contract |
| Accounting | Investment plus ongoing OPEX | Operating cost — classification with your finance team |
We are not a tax adviser. Whether a given contract is deductible operating spend depends on its structure and on your finance team. That is why we ask you to get in touch for the details — we will prepare scope, a quote and a summary you can show the CFO.
Hypervisor licence and TCO comparisons are in the VMware vs Proxmox analysis and in Proxmox implementation for businesses. This article is about something else: where the virtual machines run and how that hits the invoice.
What you get in Warsaw
A Warsaw facility is not “a rack in the office basement”. It is a professional data centre: power, cooling, physical access, connectivity. We run a Proxmox VE cluster there for your virtual machines. Being in the capital shortens the path to your office and to the carriers you already buy in Poland; the data stays in-country.
- Proxmox VE cluster — KVM virtual machines and LXC containers in one panel, with high availability on a cluster that has quorum.
- Storage and network — ZFS, Ceph or shared block sized to your profile; VLANs and isolation of production from management. Storage detail: choosing storage on Proxmox.
- A backup off the production cluster — Proxmox Backup Server as a separate backup host, with a second copy off-site or to tape. How we deliver it: ransomware-resistant backup.
- The access you choose — full UI for your team, or operator care on our side; roles and corporate login can be wired to Active Directory.
- Migration to start — from VMware, from physical servers (P2V) or from an existing Proxmox. You do not have to build a lab on site first.
We do not publish a per-vCPU or per-rack list price here. Scope (capacity, network, backup, on-call) is sized to the environment. We will audit and quote at no charge — with no obligation to move.
Who this model is for
This is for companies that want to stay on Proxmox (or move to it) but do not want to put a server room on the balance sheet. It is not a substitute for every public cloud, and it is not colocation of a bare server you wheel into a rack yourself.
| Situation | Own building | Public cloud | Proxmox in Warsaw |
|---|---|---|---|
| Steady business workloads | Full control, large CAPEX | Flexible, often more expensive on a steady load | Hypervisor control, cost as a service |
| No room for a rack | You build or rent a server room | You need no rack | You need no rack |
| Data in Poland | Yes | Depends on region and contract | Facility in Warsaw |
| Leaving VMware | A project in your own building | Often a rewrite onto the vendor’s services | Migrate to Proxmox and host in one move |
| Seasonal web peaks | Hardware sits idle the rest of the year | Cloud is often better here | For the steady core; peaks can stay outside |
If you already have a room and power, and you only need the hypervisor — stay with Proxmox implementation on your hardware. If you do not want that room at all, this model is for that.
How to start
There is no public “per VM” list price, because environments differ in network, backup and the maintenance window. The conversation looks like this:
- A short audit — how many virtual machines, what traffic, what backup RPO, whether you are leaving VMware.
- Scope and quote — capacity, network, backup, on-call; what will land on the OPEX invoice.
- A move plan — waves, windows, tests. Production does not have to wait on your hardware tender.
Proxmox Migracje is based in Warsaw and serves companies in Poland and abroad. Hosting virtual machines in a local data centre is a separate service alongside migration and implementation — get in touch for the details.
Host Proxmox virtual machines in Warsaw — without CAPEX
Tell us how many machines you have and whether you are leaving VMware. We will prepare scope, an OPEX quote and a move plan. Details after you get in touch.
⚡ Free consultation→ Proxmox implementation on your hardware