Companies are using more AI than ever and barely spending more on it, Ramp data shows
Token spend grew 1.5% across August and September after a 42% jump in July. Price cuts, cheaper models and thriftier customers are capping revenue for the big labs.
DevIsDead Newsroom
AI usage is up. AI bills are not. New data from corporate card company Ramp shows token spend nearly flat across August and September even as volumes rose, Rebecca Bellan reports for TechCrunch.
The numbers
Token spend grew just 1.5% over August and September, after jumping 42% month over month in July. October picked back up, with spend up 8.9% between September 23 and October 23, but Ramp economist Ara Kharazian says the headwinds are strong.
Nearly all AI spend comes from venture-backed, fast-growing tech companies, he says, so if they slow, they slow together. The top 1% of firms, the most AI-heavy, slipped slightly below the roughly $10,000 per employee per month they hit at their summer peak, and that alone is enough to stall overall revenue. Median firms spend $12 to $15 per employee per month; the top 10% spend $600 to $1,000. Coding agents justify big budgets in software and finance, but a doctor's office or car dealership may never need that much.
Why spending stalled
Not weak demand, Kharazian says, but competition. OpenAI halved API prices for GPT-6 Sol and GPT-6 Luna in September; Anthropic cut prices on Fable 5.1 and Opus 5.5, which it said in a filing reduced revenue growth by about $600 million. Meta, Nvidia and a crowd of open-model companies offer cheap alternatives, and tools like Jev let companies route simple yes/no work to even cheaper models.
Customers are also getting smarter: defaults that keep expensive models off long tasks, per-employee budgets, and routers that push work to lighter models. Notion co-founder Akshay Kothari summed up his own data: "tokens go brrr. costs go ahhh."
What it means
Great for buyers, harder for labs heading toward IPOs. Bloomberg reported Anthropic's potential valuation has slid to $1.2 to $1.3 trillion, and OpenAI cut its target too. The hope is a new breakthrough product, like coding agents were; Kharazian doubts personal agents will be it.
For engineering teams, this is your moment to negotiate, route aggressively and set budgets. It's a buyer's market.
Scoreboard update: Humans +1. Companies learn to use more AI while paying the labs less for it.
Stay undead.
Comments 0