EU cloud migration

Move workloads from AWS or Azure to EU providers, where it pays off

Not everything should leave the hyperscalers. Some workloads should. We tell you which ones, what it would really cost, and how to move them without a big-bang cut-over.

An EU cloud migration assessment looks at your AWS or Azure workloads and decides which could move to Hetzner or other EU providers, with a cost model, the risks and a step-by-step plan in Terraform and Ansible. It is for companies worried about cloud costs, data location or dependence on one US provider. The assessment starts at €2,500 and takes 1–2 weeks; the migration itself is quoted separately.

Who it is for

  • Companies with growing AWS or Azure bills for workloads that do not need hyperscaler features
  • EU businesses whose clients ask where their data is hosted and by whom
  • Fintechs and ICT suppliers that want a credible exit plan from a single cloud provider
  • CTOs asked by the board whether moving to an EU provider is realistic

Why now

  • The monthly cloud bill keeps growing, mainly on compute, bandwidth and storage
  • Clients or partners ask where data is stored and which providers can access it
  • DORA has applied since 17 Jan 2025, and financial-sector clients now put ICT contract clauses (Art. 28–30) on suppliers, including how you would leave a provider
  • Leadership wants less dependence on one provider, but nobody has quantified the trade-offs

What you get

  • Workload inventory: what runs where, how it is used, and which managed services it depends on
  • Cost model: current spend per workload against an EU-provider design, including the people time to run it
  • Move or stay decision per workload, with reasons: what moves easily, what should stay, and what is not worth the effort
  • Target design on Hetzner or other EU providers: network, compute, storage, backups, monitoring and access
  • Migration plan in phases, with data transfer, cut-over and rollback steps for each phase
  • Terraform and Ansible skeleton for the target, so the plan is concrete and reviewable
  • Risk list: what you lose (managed services, regions, support) and how to cover it

How it works

  1. Inventory

    Read-only access to billing and accounts; we map workloads, data and dependencies.

  2. Model

    Costs and designs for each candidate workload, including the effort to run it afterwards.

  3. Decide

    A short report with a recommendation per workload, discussed with you on a call.

  4. Plan

    A phased migration plan and infrastructure-as-code skeleton, ready to quote and execute.

Honest about what does not move

Hyperscalers are expensive for some things and very good at others. A typical AWS account mixes both:

  • Often worth moving: steady virtual machines, build runners, file storage, backups, bandwidth-heavy services, self-hosted databases, and infrastructure such as blockchain nodes that needs dedicated hardware.
  • Often worth keeping: services built on managed queues, serverless functions or proprietary databases, where moving means rewriting the application.

A good assessment can conclude that you should stay. In that case you still leave with a cost map, a cleaner account and an exit plan you can show clients.

How we move what does move

Every target environment is written in Terraform and configured with Ansible before any data moves. Each workload moves on its own: we copy data, run both sides in parallel where possible, switch traffic, and keep the old side ready for rollback until you approve. Monitoring and backups are in place on the new side before the cut-over, not after.

What we do not do

We do not push a migration that does not pay off, and we do not move everything at once. If the numbers say stay, the report says stay.

Related work

Questions clients ask

Should we leave AWS completely?

Usually not. Full exits are rare and often not worth it. The usual result is a partial move: steady compute, storage, build runners or bandwidth-heavy services go to an EU provider, while workloads that depend on managed AWS or Azure services stay. The assessment tells you where your line is.

How much can we save by moving to Hetzner?

It depends on the workload, and we will not quote a percentage before seeing yours. Steady compute and bandwidth often cost much less on dedicated or EU cloud servers. Managed databases, queues and serverless are harder to replace and need people time to run. The cost model shows both sides.

How much does the assessment cost?

From €2,500 excluding VAT, usually in 1–2 weeks. The migration is then a separate fixed-scope quote based on the plan.

How long does a migration take?

It depends on how many workloads move and how much data. We plan it in phases, one workload at a time, each with its own cut-over and rollback, so the business keeps running throughout.

Which EU providers do you work with?

Mostly Hetzner, including dedicated servers, plus other EU cloud and hosting providers when they fit better. We describe everything in Terraform and Ansible, so you are not locked into us or the provider.

Does this help with DORA or data-location questions?

It can. A documented exit plan and a clear map of where data is hosted are things financial-sector clients ask for. The assessment gives you both, even if you decide to stay.

Find out what is worth moving

Tell us roughly what you run and what you pay. We will say honestly whether an assessment is worth it for you.

Talk to an engineer