compute inflation
/kəm-PYOOT in-FLAY-shən/·noun
The sustained rise in the cost of computing power — and its pass-through into the price of everything computing touches.
Definition
Economics. A broad upward pressure on prices caused by surging demand for computation: chips, memory, servers, cloud capacity and the energy to run them. As AI absorbs semiconductor and datacenter supply, costs propagate into consumer electronics, software subscriptions and energy bills.
Business. The rising unit cost of an organization's computational work — cost per workload, per query, per token — even as hardware nominally improves, because demand outruns supply.
Origin
First recorded 2026, during the AI infrastructure boom, when analysts began describing chip scarcity, datacenter capex and electricity demand as a single inflationary force rather than separate stories.
Usage
- Compute inflation is the tax nobody voted for: it arrives in your laptop's price tag and your power bill at the same time.
- The board approved the AI roadmap but not the compute inflation that came with it.
- Analysts warn that compute inflation could outlast ordinary inflation.
Frequently asked
Is compute inflation the same as regular inflation?
No. Regular inflation measures a broad basket of goods. Compute inflation names one specific driver inside it: the rising cost of computation and its knock-on effects on electronics, cloud services and energy prices.
What causes compute inflation?
Demand for AI training and inference growing faster than the supply of advanced chips, memory, datacenter capacity and grid power.
Who feels compute inflation first?
Gamers and PC builders (hardware prices), startups (cloud and API bills), and households near datacenter clusters (electricity rates).
Can compute inflation reverse?
If chip supply, efficiency gains or an AI demand slowdown outpace consumption, unit costs can fall — as happened in past semiconductor cycles.