This is running on a single AWS c5.metal instance. It has 196GB RAM, but we are using only 40GB at peak. The instance provides two 24-core CPUs, but we are only using one of them on 23-threads. So far, it's handling the flow of connections and requests from HN fine. We over provisioned just in case as we were not sure what to expect :-)
Storage is an AWS EBS volume.