Most of the infrastructure we walk into looks the same: a VMware environment that's been running for years, quietly getting more expensive, and harder to change. The ambition to do more — to automate, to adopt AI, to move faster — is usually there. The infrastructure to support it often isn't.
VMware environments tend to be stable and well-understood — that's exactly what makes them hard to leave. But the cost keeps climbing. Licensing, support, and the operational weight of a platform you can't easily modify add up. Money that could go into building new things keeps going into maintaining old ones.
A Practical Alternative: Proxmox and Hybrid Flexibility
Our starting point is usually the hardware that's already there. We replace the virtualization layer with Proxmox VE, extend into hybrid cloud where it makes sense — Hetzner is a reliable and cost-effective option — and open the door to modern workloads: containers, AI pipelines, whatever the business actually needs.
The result is an infrastructure that costs less to run and is easier to grow.
Proxmox has earned its place in production environments. It's not a hobbyist tool or a compromise — it handles demanding workloads, has a strong community, and is actively developed. Teams coming from VMware find the concepts familiar. For anyone already comfortable with Linux, it feels like home.
Unlocking Local AI with Hardware Passthrough
One capability that often surprises people: hardware passthrough. With Proxmox, you can assign a GPU or specialised AI accelerator directly to a virtual machine — the VM gets bare-metal access to the hardware.
This means you can run local LLMs, document processing pipelines, or invoice extraction workflows entirely within your own environment. No external APIs, no data leaving your network.
We've built these setups for clients. If you have the hardware, we can help you put it to work.
Why Now Is the Right Time to Switch
VMware's licensing model has changed significantly in recent years, and not in customers' favour. For many organisations, the cost of staying put now exceeds the cost of migrating.
What's different today is that the alternative is genuinely ready. Proxmox isn't a risk you take to save money — it's a solid platform that also happens to cost less. And in many cases, the savings from dropping VMware can pay for the migration itself.
A Typical Customer Story
Take a mid-sized manufacturing company — we'll call them Northwind Components.
They had a solid VMware environment, an experienced team, and systems that ran without drama. But behind the scenes:
- Licensing and support costs were going up every year
- Hardware was aging and due for replacement
- There was no budget left for anything new, let alone AI
Every euro was tied up keeping the lights on.
The Hidden Problem: Success
Ironically, their biggest challenge was that their system worked too well.
Their infrastructure had been optimised to the limit — every server fully utilised, every resource accounted for. Like an overloaded camel, it had no spare capacity. Even attempting to migrate a single workload risked destabilising the whole environment.
This is a common pattern: the technical challenge isn't the migration — it's the lack of breathing room to begin.
How We Solved It
The solution was to stop trying to migrate within a full environment. Instead, we used the projected VMware savings to temporarily expand capacity:
- Rented a few dedicated servers from Hetzner
- Stood up Proxmox alongside the existing stack
- Started moving workloads across one by one
That breathing room made all the difference. As each workload moved over, licensing costs dropped and the pace picked up. By the end, the project had effectively paid for itself.
The Outcome
Once the migration was complete:
- Infrastructure costs dropped significantly
- Performance improved
- No more vendor lock-in
- Room to actually invest in AI and automation
In their own words, "it felt like we could breathe again."
That's usually the moment we're working towards.
More Than Migration — A Foundation for Innovation
The migration is the starting point, not the finish line.
Once the infrastructure is lean and under your control, things that felt out of reach become straightforward — running local LLMs, containerising applications, experimenting without asking for budget approval every time.
We help with the transition, but what we're really building is the foundation for what comes next.
Let's Build the Next Layer Together
If any of this sounds familiar — the stable system that's quietly getting harder to justify, the AI projects that keep getting pushed back — we'd be happy to talk.
We can usually tell pretty quickly whether a migration makes sense for your situation, and what it would realistically involve.