
A standardized economic asset representing GPU-hours, accelerator-hours, FLOPS, tokens/second of measurable ai computing infrastructure capability under defined operating conditions.
GPU-hours, accelerator-hours, FLOPS, tokens/second define the canonical capacity units.
Metered gpu-hours, jobs completed, tokens processed, and reserved capacity establish the operational record.
Scarcity, utility, durability, demand, transferability, and revenue potential determine institutional relevance.