I can speak a bit to this. I previously lead the database team at Heap where we had one of the largest deployments of Citus with ~1PB of data. Citus and Aurora

I can speak a bit to this. I previously lead the database team at Heap where we had one of the largest deployments of Citus with ~1PB of data. Citus and Aurora Serverless are pretty different. The only thing in common is that they are both related to scaling Postgres. Citus makes it easy to horizontally scale Postgres. It let's you take a cluster of Postgres instances and make them appear as a single Postgres instance. This let's you scale Postgres to petabyte scale datasets (and beyond) because under the hood you are storing the data across multiple servers. On the other hand, AIUI, the benefit of Aurora Serverless is the auto-scaling. It will automatically add or remove compute on an as-needed basis. This means you no longer have to worry about provisioning a Postgres instance of the right size. You can also save money depending on your usage pattern. Even though it will scale Postgres, you are still limited to a maximum DB size of 32TB. Aurora Serverless is probably a good fit if you are under this size, but once you pass 32TB, you will need something like Citus to scale further. As for extensions, you are limited to the extensions Amazon provides. As you point out, Aurora is mostly at the storage layer. For that reason I think they could add most extensions if they wanted to. Each extension brings on some risk to them so they audit every extension before approving them. I've heard issues with extensions are one of the biggest sources of support requests.