European Internal Developer Platform (EIDP) - Cost comparison: US vs EU clouds
Source:
- https://eidp.com/cloud-blog/
- https://www.linkedin.com/posts/steinwelberg_sovereignty-independency-cloudcosts-activity-7424345942071455744-SOC7/
Yeah, sorry for the LinkedIn post, but from it,
In a short while I will be writing a blog post on how I got to these percentages.
13 replies
Do any of these providers offer anything like Lambda, S3, Dynamo and EventBridge?
Scaleway has pretty much all one needs to compete with AWS.
Interesting. Just had a look at scaleway and can see it has functions, and surprisingly even an object storage service called... S3. How are they getting away with that?
I can't see any dotnet examples for functions but maybe I haven't looked enough. Certainly looks interesting though, I'll look further!
Oh it has a terraform provider too, nice!
S3 has become an API that can be implemented by different software and providers
https://garagehq.deuxfleurs.fr/
https://cloud.ionos.com/storage/object-storage
https://www.scaleway.com/en/object-storage/
https://blocksandfiles.com/2025/04/14/storadera-s3-cloud/
I'm going to look into migrating to Scaleway, looks to have the serverless stuff I'd need!
Let us know how it goes!
Yes
Yes, but I still dont see any point of lambdas.
At the end it generates more costs than just simple autoscaling
The two huge benefits are that you only pay for what you use and you don't need to spend any effort managing it. For my own personal stuff, it makes zero sense to have something running continuously as the number of invocations per week is like less than 10!
The time spent managing stuff can't be understated either. At my prevous job, we used EC2, and involved a fair bit of management: always doing deploys to roll out the latest AMI's with security patches, doing load tests for different trading peaks and working out how many instances we'll need at what size, monitoring the instances during deployments on the chance that an instance might not be brought back in, which would sometimes happen.
My current job, it's 100% serverless. The website, any API's, everything. We never have to think about scaling or patching or load testing. The running costs maybe higher but the actual costs for the teams is less.
My latest project at work was recoding serverless to just ordinary code, because of the costs.
You can always have containers, autoscaling with k8s….
Came here to ask this, The answer has to be "No"
Turns out I was (at least partially) wrong. Scaleway offers serverless functions !
Why do you think customers choose to pay a huge premium for AWS? Who wouldn't choose an 86% percent saving? Are they dumb?
Understand why they aren't choose Hetzner and talk about that, not oversimplifying raw compute metrics and completely missing the mark on the differentiators.