# Forecasting AI spend with cohort modeling

> Linear extrapolation is why your end-of-month bill keeps surprising you. Cohort-based forecasts predict the next invoice within ±3%.

- Source: https://spendtensor.com/blog/forecasting-ai-spend-cohorts
- Author: Spend Tensor
- Published: May 28, 2026
- Category: Research
- Reading time: 7 min
- Topics: Forecasting, Data Science

---

If you forecast AI spend by extrapolating last week's daily run rate, you will be wrong every month. Usage doesn't grow linearly — it grows by feature cohort, and each cohort has its own adoption curve.

Our forecasting model treats every feature × tenant pair as a cohort. We fit a logistic curve to early usage, weight by recent traffic, and roll up to a portfolio-level estimate. Across our customer base, the median forecast error for the next 30-day invoice is 2.8%.

---

Published by SpendTensor — AI spend management and LLM FinOps. Quote with attribution and a link to https://spendtensor.com/blog/forecasting-ai-spend-cohorts.