Move to EU providers
Leave AWS and the other US clouds for European providers, without breaking anything.
More teams want their product off US infrastructure: because the bill is too high, because their customers ask where the data lives, or simply because they would rather not depend on US companies. We plan and do the move to European cloud providers such as Hetzner, Scaleway and OVHcloud, part by part, with your product running the whole time.
One client, start to finish
We can't name them, but the numbers are theirs. They came in paying about $3,000 a month on AWS. The first round never touched a server: we fixed what the app was doing, and the bill fell to about $2,000. They had been wanting off US providers anyway, so we moved the infrastructure to European hosts next, which brought it to about $1,200. The last few services that did not move cleanly now run as self-hosted open-source software on one standalone server, and the monthly bill sits between $600 and $800.
That order matters. Cut first, and you know exactly what you use before you pick a new home for it. Move first, and you carry the waste with you.
Where things usually end up
For plain compute we usually start with Hetzner (Germany): cloud servers for most apps, dedicated machines when the database is big. When a team wants more managed pieces, Scaleway and OVHcloud (both France) have managed Postgres, object storage and serverless functions, so Lambda-style code has somewhere to go.
S3 moves to Hetzner or Scaleway object storage with the same API, so the code change is mostly an endpoint and a key. CloudFront becomes Bunny.net (Slovenia). Vercel and Netlify sites can deploy to your own server with Coolify. Supabase can be self-hosted on an EU server and keeps the same API your app already calls. Email from SES or SendGrid goes to Brevo or Mailjet.
None of that is a rule. A product with three users and a cron job needs one small server; a product with real traffic and a big database needs more care, and sometimes the honest answer is to leave a piece where it is.
How the move goes
We map what you run and what each part costs; if you have done a bill review with us, half of that is already done. Then we build the new setup next to the old one, copy the data over and keep it in sync, and send real traffic to it before anything is switched off. The cutover is scheduled with you, and the US services are closed only after the new setup has run clean for a while.
Questions
Is it actually cheaper?
Usually, and by a lot on raw servers and bandwidth. A dedicated Hetzner box with 64 GB of RAM costs less per month than a mid-size RDS instance, and outbound traffic is included up to generous limits instead of billed per GB. What you give up is the long tail of managed services, so a bit more of the setup becomes yours. We tell you up front which parts are worth moving.
Do I have to move everything?
No. Plenty of teams move the servers and the database and leave email or a niche managed service where it is, because moving it would not pay off.
What about downtime?
For most apps the only real downtime is the database switch. We keep the new database in sync with the old one, pick a quiet hour with you, pause writes briefly, switch over, and keep the old setup ready to switch back to for a while.
What does the move cost?
We quote it after we have looked at your setup, before any work starts. If you want the current bill cut first, that part runs on the usual deal: we are paid from the savings.