AFAIK Google deploys massive HDDs to get massive parallel I/O. Possibly they changed storage strategy to mainly use SSD (lower capacity, higher IOPS) , so they

AFAIK Google deploys massive HDDs to get massive parallel I/O. Possibly they changed storage strategy to mainly use SSD (lower capacity, higher IOPS) , so they no longer able to provide wasting capacities?