AI companies are now racing to the bottom — crashing token prices and competitive models push companies to cut costs
TL;DR
All major AI developers are cutting prices to keep pace with capable new model releases from China. The competition is pushing down cost per token and therefore revenue per request. At the same time, spending on data centers, chips, and model training remains high. That makes the profits underpinning current valuations and financing plans harder to reach.
Nauti's Take
The price war is a real opportunity for small teams: workflows that failed on token costs a year ago now pencil out. The risk sits on the other side of that math, since shrinking margins mean unstable pricing, shifting rate limits, and models that get retired on short notice.
Teams building on this should measure cost per completed task rather than cost per token, and plan for a provider switch from day one.
Sources