Ramp data highlights a gap between AI usage and spending
New reporting examines spending patterns in a defined customer sample, rather than the entire AI market.
Using more AI does not always mean paying proportionately more for it.
What happened
TechCrunch reported on 6 November that Ramp's data showed much slower token-spending growth during August and September than in July, followed by renewed growth in October. The analysis draws on customers tracked through Ramp's spending products.
Why it matters
Lower effective prices and customers choosing cheaper models can let usage rise faster than bills. That matters for software buyers managing costs and for model providers trying to turn growing activity into revenue.
The bigger picture
This is evidence from a particular sample and historical period, rather than a complete measure of the AI economy. The underlying data predates the article; it should not be presented as November usage or proof that AI demand has stopped growing. Customer mix, prices and the measurement window all affect the interpretation.