Compute
FoundationsAI Economics and Infrastructure
Also called: Computing Power
Compute is the raw hardware processing power, measured in floating-point operations per second (FLOPS), required to train, fine-tune and run artificial intelligence models. It runs on specialised processors such as GPUs and TPUs. Compute is the primary capital cost of AI, the bottleneck on how fast a system can be scaled, and the strategic resource everything else depends on.
In practice
Compute is the dominant operational cost line in software margin analysis and technology budgeting. A business that buys AI rather than builds it meets that cost as inference pricing, which grows with usage rather than headcount. Model the bill at ten times current volume before signing, because that is where the margin question is decided.