For us, we only allocate ie 8 runners to a VM with 8 cores, which we pin to 8 cores on the actual server, so our focus is getting as many cores as possible to g

For us, we only allocate ie 8 runners to a VM with 8 cores, which we pin to 8 cores on the actual server, so our focus is getting as many cores as possible to get as many runners as possible. We kind of expected that scaling HW with more CPUs instead of more entire rack systems would be more energy efficient, but I don’t think it makes too much of a difference compared to other factors like CPU/vendor architecture. FWIW our energy bill is $1000+ per month just from servers, and then several $k for various licensing on each host per year, so there was some sweet spot we were optimising for (low power, but less licensing cost via less racks). Our cost in the cloud would be $5k per month or more, so either way we still saved money this year.