c6a.16xlarge

Analysis

The c6a.16xlarge is a compute optimized instance in the c6a series with 64 vCPUs and 128 GiB of memory on AMD EPYC 7R13 (up to 3.6 GHz), which makes it compute-heavy at 2 GiB of memory per vCPU.

It is 0.6% cheaper than the older c5a.16xlarge on AMD EPYC 7R32, so staying on the previous generation costs more as well as being slower. The newer c7a.16xlarge on AMD EPYC 9R14 is 34% dearer at $3.28/hour.

Running it in Mumbai instead of São Paulo saves $1,662 a month per instance: $1.50/hour against $3.77/hour, a 152% spread.

Within the same generation and size, the c6g.16xlarge on AWS Graviton2 costs $2.18/hour, 11% less ($199 a month), though switching means rebuilding software for arm64. The other variants are the c6gd.16xlarge at $2.46, the c6i.16xlarge at $2.72 and the c6gn.16xlarge at $2.76.

At $2.45/hour on demand in US East (N. Virginia), about $1,787 a month, the c6a.16xlarge is one of the cheaper ways to buy compute optimized capacity: only 8% of the 291 current compute optimized types cost less per vCPU-hour than its $0.0382.

Reserving it for a year with no upfront payment cuts the Linux rate to $1.62/hour, which beats on demand once it runs more than 66% of the time; three years brings it to $1.08/hour, 56% below on demand. Windows Server roughly doubles the bill at $5.39/hour, because the licence is charged per vCPU.