AI
Ramp data shows slower AI-spending growth in August, not an economy-wide retreat
TechCrunch reports that 56% of companies in Ramp's 70,000-customer spending dataset paid for AI products in August, with adoption growing only slightly from July. AI spending per employee in the highest-spending 1% of firms fell nearly 10% to $7,205. Lower token prices and summer seasonality complicate the interpretation: declining spend need not mean declining use, and Ramp's technology-heavy customers are not representative of all businesses.
Editorial summary · Publisher article
What matters
- The reported average token cost fell from $1.15 per million in March to $0.68, reducing spend for a given workload.
- A Census survey cited in the article shows much lower business adoption, illustrating the importance of the sample.
- Only 6.4% of AI-spending companies in Ramp's sample used model-serving or inference platforms in August.
The revenue question
For model providers, price reductions require enough additional usage to compensate. For customers, cheaper models and lower token bills can be beneficial. One month of payment data cannot separate vacations, substitution toward cheaper models and a durable change in demand; Ramp saw a similar late-summer slowdown the previous year before growth resumed.
Context & caveats
- These are Ramp-customer spending observations, not audited financial results for AI labs or a representative economy-wide adoption measure.
Why this made the edition
Complete supplied article read for the September 9 morning refresh. Attribution, timing and uncertainty retained; compared with previous editions and other selected coverage.
Source material