c8gb.12xlarge

Analysis

The c8gb.12xlarge is a compute optimized instance in the c8gb series with 48 vCPUs and 96 GiB of memory on AWS Graviton4 (up to 2.8 GHz), which makes it compute-heavy at 2 GiB of memory per vCPU.

Within the same generation and size, the c8g.12xlarge on AWS Graviton4 costs $1.91/hour, 33% less ($679 a month). The other variants are the c8i-flex.12xlarge at $2.14, the c8i.12xlarge at $2.25 and the c8gd.12xlarge at $2.35.

Running it in N. Virginia instead of Frankfurt saves $288 a month per instance: $2.84/hour against $3.24/hour, a 14% spread. It is not offered in Canada Central, São Paulo, London, Mumbai, Singapore, Sydney and Tokyo, so plan around that if you need a presence there.

Per vCPU it is one of the dearer compute optimized types: 82% of the 291 current types in the family cost less than its $0.0592 per vCPU-hour. The premium mostly pays for the very fast networking. On demand it is $2.84/hour in US East (N. Virginia), or $2,076 for a 730-hour month.

Looking beyond its own line, 24 of 30 current types with 48 vCPUs and similar memory are cheaper. The cheapest of those is the c6g.12xlarge on AWS Graviton2 at $1.63/hour.

It is size 7 of 10 in the c8gb series. Stepping up to the c8gb.16xlarge gives 64 vCPUs and 128 GiB for $692 more a month. Dropping to the c8gb.8xlarge saves $692 a month.