What it means
Cost per token is the all-in price to generate a unit of model output — and it is the point where the entire AI stack converges. Every layer beneath it feeds into this number: the price of accelerators, how efficiently they run inference, the memory and networking around them, and the electricity powering it all. Driving cost per token down is what makes new AI products economically viable, so it is the metric operators relentlessly optimise. Because it rolls up the whole chain into one figure, it is also the cleanest way to see whether AI is getting cheaper to run — and therefore more widely deployable.
Why it matters to investors
Cost per token is the bottom line of the AI supply chain — the single number that compresses chips, memory, networking and power into one. Falling cost per token expands what AI can profitably do, driving more usage; investors can read it as the convergence metric that links every layer of the chain to actual economics.
Companies on this part of the chain
Named to show where the term sits in the AI supply chain — research, not advice, and never a recommendation to buy or sell.
Related terms
See Cost per token in the live AI chain.
THE ENTITY maps every constraint onto one live model — which part is tight now, who owns it, and who gets squeezed when it moves. Plain-English reads you can check.
THE ENTITY is an educational read on the AI supply chain — research, not investment advice. It explains how the chain works and who sits where, never price targets or buy/sell calls.