Ask HN: Why does Azure seem so resource-constrained – especially with GPUs?

I've been struggling to get any GPU-powered-family Azure resource quotas increased despite following the directions explicitly and filing the requisite support tickets; it's apparently a "bi-monthly" check-in that will happen now that I've been denied. Is there a global squeeze on GPUs, and it's reserved for OpenAI, or am I doing something base-line wrong, perhaps?

Going to go spend some time on Google Cloud and see if I can get an Unreal Engine capable VDI up.