The Kobeissi Letter cites Ramp data showing monthly AI token spend growth slowed to 0.2 percent in September
BREAKING: AI token spending grew just +0.2% MoM in September, the lowest monthly growth since November 2025. This follows a +1.0% increase in August, marking two consecutive months of growth at or below +1.0%, the weakest streak in at least 18 months. This comes as falling prices for the most advanced AI models, alongside improvements in efficiency and quality, drive down the average price per million tokens processed by these models. Since July, overall AI token prices have declined -48%, to $0.65 per million tokens, the lowest since at least January 2026. Lower costs could unlock the next wave of AI adoption.
Context
A sharp deceleration in month over month corporate AI token spending on Ramp highlights how falling per token API prices, model routing, and post summer budget discipline can flatten dollar spend even while raw token volume grows.
Corporate spend management platform Ramp tracks card and invoice payments across thousands of US businesses, offering a window into enterprise software and API outlays. Throughout 2025 and the first seven months of 2026, Ramp's monthly AI token spend series frequently recorded 20 percent to 58 percent month over month jumps.
Deep dive
On October 23, market commentary account The Kobeissi Letter published a chart citing corporate spend platform Ramp that tracks the month over month percentage change in AI token spending from February 2025 through September 2026. According to the chart, monthly corporate token spend growth slowed to 1.0 percent in August 2026 and 0.2 percent in September 2026, marking the slowest monthly expansion since a 2.0 percent contraction in November 2025.
Earlier in the series, Ramp's tracked token spending surged 58 percent month over month in October 2025, 49 percent in February 2026, 45 percent in March 2026, and 42 percent in July 2026, resulting in a 22 percent compound monthly growth rate across the full 20 month window. The sudden flattening in August and September therefore represents a sharp break from the rapid dollar expansion seen during spring and mid summer.
Because Ramp measures corporate dollar billing on its card and bill pay platform rather than raw token volume across the entire industry, several factors can compress monthly dollar growth simultaneously. Aggressive API price cuts across frontier and open weight models, enterprise migration from metered per token billing to annual committed contracts, and prompt caching all reduce incremental monthly card charges even when underlying token consumption on routers like OpenRouter continues climbing.
What to expect next
Analysts watching fourth quarter AI revenue will look to separate unit price deflation, prompt caching efficiencies, and enterprise contract shifts from genuine demand saturation as frontier labs release new reasoning and coding models, open weight alternatives narrow the capability gap on routine production tasks, and corporate procurement teams finalize their annual 2027 software and cloud compute budgets across both direct vendor contracts and API routers.