David Heinemeier Hansson

Our cloud spend (sans-S3) is down by 60% already. From around $180,000/month to less than $80,000. That's a cool million dollars in savings at the yearly run rate. The only way to get pricing in the cloud down from obscene to merely offensive is through reserved instances.

Our cloud exit has already yielded $1m

I remember when I had all our infrastructure in two datacenters which was more than $1.5M equipments. back then we had 2 full time ops and IT companies to handle hundreds of servers in our data center.
I ended up moving all our services to Google Cloud even though AWS had the majority of market share and and my team even had more experience with AWS. But services like BigQuery and BigTable in GCP could not only save us more than million dollar over a year but also help us to process data 100s of times faster!
Over the last 5 years I helped data companies move out of AWS to GCP to save millions of dollars.
With all that said, I wonder if you look into other alternatives such as GCP and leveraging services like BigQuery, BigTable, and Spanner before taking everything back to datacenter and on your own?
Lets say you want to do AI-project with tons of GPU/TPU and storage, how you are planning to do that on own servers? How many people need to manage your data centers, and how you plan updating servers and handle disk and server failures?
I personally can not imagine going back to my own servers and dealing with data centers anymore.

How much did the engineering work cost for this move though?

This is amazing. Build scale your app so you can easily move away from the cloud