United companies are paying less for artificial intelligence even as they use more of it, according to a recent study released by Ramp's AI Index, which tracks spending and usage across a large sample of American businesses. The findings indicate that the ongoing AI price wars are not hurting demand. Instead, falling prices are making it easier for companies to deploy AI at scale. If the trend continues, firms that supply chips, cloud capacity, and inference stand to see a corresponding boost to their financial results.

Usage up 50%, spending down since July

Ramp economist Ara Kharazian said the amount of money businesses inject into AI has dropped after hitting a record high in July, when the market was wide open for providers such as OpenAI and Anthropic to lower their prices.

Usage has moved in the opposite direction. Through September, usage rose nearly 50% and hit an all-time high. Ramp's usage figures are counted in tokens, the basic units of text that AI models process. In the final week of the month, Anthropic's token usage stood at 51%, while OpenAI's was 44.5%, according to Ramp, which noted that open-source alternatives remained below 5%. Together, the two providers accounted for more than 95% of the token activity in Ramp's sample.

Ramp's figures draw on transaction data from 70,000 U.S. firms, though its token data comes from a smaller, API-focused sample that prioritizes large purchasers.

According to Kharazian, competition between Anthropic and OpenAI is “making AI more accessible” to more people while driving down costs for companies.

Why cheaper tokens can mean more AI, not less

Lower prices do not necessarily imply a declining AI industry. Frank Flight, a specialist at Citadel Securities, argued in his tokenomics evaluation that the degree of adoption is increasingly reliant on the affordability and availability of compute, power, and inference capacity.

Ramp's own data reflects the price slide: the effective token price decreased by 41%, to about $0.68 per million tokens, according to the AI Index's blended token price tracker.

Enterprises route work to the cheapest model that fits

Companies are also becoming more selective about which models handle which tasks. The State of Tokenomics survey of 472 companies found that 86% were using or evaluating model routers, and router users were four times more likely to show measurable value to CFOs. The pattern suggests enterprises are matching models to workloads rather than relying on frontier models for every task.

Reliance on frontier models is also loosening. Today, 51% of respondents say they depend heavily on them, while only 24% expect to still be in that category in a year.

Tighter cost controls, but still more deployment

KPMG's Q3 AI Pulse found nearly six in ten leaders reporting measurable AI value, led by productivity at 55%, followed by faster decision-making at 49% and stronger financial performance at 37%.

“AI's value story is getting sharper,” said Todd Lohr, KPMG's Vice Chair and Head of Client Technology & Innovation.

BCG likewise found that almost half of companies now generate value from AI. Taken together, the surveys point to companies exercising tighter cost discipline even as they expand AI deployment.

What falling prices could mean for the market

Cheaper AI could bring automation into workflows that were previously too expensive to automate. The OECD's AI markets study found quality-adjusted language-model prices falling sharply, while noting that AI agents can consume far more tokens per task.

That means cheaper tokens may lower unit costs without reducing total AI bills. As adoption spreads, companies may simply use much more AI.

Source: Cryptopolitan