m7a.16xlarge

Analysis

The m7a.16xlarge is a general purpose instance in the m7a series with 64 vCPUs and 256 GiB of memory on AMD EPYC 9R14 (up to Up to 3.7 GHz), which makes it balanced at 4 GiB of memory per vCPU.

Within the same generation and size, the m7g.16xlarge on AWS Graviton3 costs $2.61/hour, 30% less ($802 a month), though switching means rebuilding software for arm64. The other variants are the m7i-flex.16xlarge at $3.06, the m7i.16xlarge at $3.23 and the m7gd.16xlarge at $3.42.

Compared with the previous generation, it costs 34% more than the m6a.16xlarge on AMD EPYC 7R13 ($690 more a month per instance), in exchange for the newer processor. The newer m8a.16xlarge on AMD EPYC 9R45 is 5.0% dearer at $3.90/hour.

Running it in N. Virginia instead of Tokyo saves $790 a month per instance: $3.71/hour against $4.79/hour, a 29% spread. It is not offered in Canada Central, São Paulo, Mumbai and Singapore, so plan around that if you need a presence there.

On demand it costs $3.71/hour in US East (N. Virginia), $2,708 a month if left running, which puts it mid-table for general purpose types at the 54th percentile per vCPU ($0.0580 per vCPU-hour) and the 54th percentile per GiB ($0.0145 per GiB-hour).