Brief IA
›

United States: Record AI Spending in October, Up 1.65%

Models & LLM·Tom Levy·

United States: Record AI Spending in October, Up 1.65%

United States: Record AI Spending in October, Up 1.65%
Key Takeaways
1Spending on AI by American companies increased by 1.65% in October, reaching a new high
2Budgets are shifting towards customer support, with its share rising from 0.8% to 13.3% in four months
3The average price per million tokens has dropped by 21%, while 1% of companies generate 80% of the revenue for OpenAI and Anthropic
Why it matters : These figures illustrate a slowed yet persistent growth, a redistribution of usage, and a strong concentration of revenue, based on data with limited representativeness.

American companies' spending on AI reached a peak in October with a monthly increase of 1.65%. At the same time, the price per million tokens has dropped significantly, and usage is shifting towards customer support. Revenues remain concentrated among an elite group of users, while the data comes from customer payments to Ramp with declared limits.

Declining Costs and Pressure on Margins for Providers

The average price per million text tokens fell by 21% last month. Ramp indicates that usage volume is growing faster than spending, contributing to the decline in costs, driven by competition between OpenAI and Anthropic and the arrival of open-weight models from China. The share of models considered the most expensive slightly increased in October, reaching 69.5% of total spending. Ara Kharazian cites Opus, Fable, Sol, and Astra in this category. Despite this, observed use cases often favor faster and cheaper models over the more expensive flagship offerings. Ramp estimates that the arrival of new clients could reduce margins for OpenAI and Anthropic.

Who Pays: Spending Concentrated Among a Minority

One percent of companies generate 80% of the revenue for OpenAI and Anthropic. Ramp describes this first circle as primarily composed of venture-backed companies in technology, finance, and technical services, while AI adoption struggles to progress outside these sectors. Among the biggest spenders, the median reaches $10,200 per employee per month, up 9%. In contrast, a typical company spends $16.32 per employee per month, with an increase of less than a dollar over the month.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Budget Reallocation: Customer Support Accelerates

In October, software development teams accounted for 72% of tracked spending, but their budgets have decreased from summer highs. Customer support and success services are showing the fastest growth, with their share rising from 0.8% in June to 13.3% in October, likely driven by the use of AI chatbots in support. Leadership teams are also slightly increasing their spending. Ara Kharazian notes, however, that the rise in customer support does not match the pace observed earlier in the year in development.

OpenAI Grows Faster than Anthropic, with Data to Handle with Caution

Commercial client spending for OpenAI increased by 2.1% month-over-month, compared to 0.5% for Anthropic. Ara Kharazian points out that already identified structural risks persist. Overall, AI spending in the United States reached a new high in October with a month-over-month increase of 1.65%. These figures come from Ramp's AI Index, based on payments from this provider's enterprise clients, which aggregates data from over 85,000 U.S. companies. Ramp specifies that free tools and usage through personal accounts may escape tracking. The separate metrics on token volumes and spending only cover companies that have activated Ramp Token Spend Management and do not reflect the entirety of the platform. Ramp also describes a larger, tech-savvy, growth-oriented clientele, which likely leads to an overestimation of adoption.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.