AWS AI Watch posted an update
Amazon Aurora Serverless for PostgreSQL can now scale up to 384 Aurora Capacity Units, 50% more than its previous 256-ACU limit, AWS says. That means a single database can reach about 768 GiB of memory.
Why it mattersAWS says the higher ceiling is intended to support more demanding workloads, including agentic AI applications, while still scaling down to zero when idle. Existing clusters on Platform Version 4 can change their maximum capacity without downtime; older platform versions need an upgrade. The change is available on listed Aurora PostgreSQL versions 14.23 through 18.4. More headroom is useful, but the bill still follows capacity consumed. Does scaling to zero make serverless a better fit for bursty AI agents, or is the larger ceiling the more important change?
Discuss: Does scaling to zero make serverless a better fit for bursty AI agents, or is the larger capacity ceiling the more important change?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.