Not necessarily compute minutes but ancillary costs like egress and especially local block storage. I recently tested a training cluster and these costs were n

Not necessarily compute minutes but ancillary costs like egress and especially local block storage. I recently tested a training cluster and these costs were nearly as much as the GPUs. And they’re both extremely high margin as well as lower QoS than what you can get in bare metal.