DEMAND
Data centers
GPU racks, cloud ASICs and general-purpose servers drive the category.
NVIDIA
GB300 to Vera Rubin: memory per rack
The conversion is capacity per rack in TB multiplied by rack volume, divided by 1,000,000. Annual totals and year-on-year increments remain separate.
CUSTOM ASIC
Non-NVIDIA ASICs add 2.705 EB
The base model for Google, AWS, Meta, Microsoft, OpenAI and other ASICs adds 2.705 EB. Combined with NVIDIA, the strictly additive increment is 5.839 EB.