AI Business Spending Flatlines as Model Prices Fall, Ramp Finds
A new report from Ramp reveals business AI spending flatlined in late summer despite rising usage, as model price cuts and efficiency gains lower overall costs.
In late summer, corporate spending on artificial intelligence stalled even as actual usage continued to climb, according to new corporate card data from financial platform Ramp. Between August and September, growth in token spending dropped to just 1.5%, down sharply from a 42% monthly surge recorded in July. Ramp lead economist Ara Kharazian noted that the spending plateau is driven by intense price competition among major developers like OpenAI and Anthropic, alongside smarter cost management by business customers.
Why is AI spending slowing down?
The report indicates that while token volume—the measure of text processed by AI systems—is rising, overall costs for users are falling. In September, OpenAI cut API pricing in half for its GPT-6 Sol and GPT-6 Luna models. Anthropic similarly reduced rates for its Fable 5.1 and Opus 5.5 models, a decision that the company stated in a filing slashed its revenue growth by roughly $600 million.
Competition from alternative providers is also keeping prices low. Companies like Meta and Nvidia are offering enterprise options with low-cost or open-source models. Organizations are also adopting lightweight tools such as Jev to route simple yes-or-no queries to the cheapest available models.
Furthermore, corporate buyers are becoming more disciplined in how they deploy AI automation. Rather than allowing unrestricted access to premium models, companies are setting employee budgets, applying default settings that prevent long-running tasks on high-end models, and using routers to shift basic tasks to light models.
What it means for small businesses
For small business owners, the shift toward a buyer's market means that powerful AI capabilities are becoming far more affordable. Price cuts from major model providers allow businesses to deploy AI customer service, voice tools, and workflow automation without facing runaway software expenses.
Ramp's figures show that typical business spending on AI remains quite modest. Median companies spend between $12 and $15 per employee per month on AI tools, while the top 10% of spending firms spend between $600 and $1,000 per employee. Only a tiny fraction of high-growth tech firms spent up to $10,000 per employee at their peak in early summer, a level that economist Kharazian notes typical businesses like doctor's offices or car dealerships do not need to reach.
Small business owners can keep their AI automation costs down by following the practical steps larger firms now use: selecting cheaper, specialized models for simple tasks, establishing clear usage budgets for staff, and taking advantage of recent provider price cuts.
How model providers are impacted
Slower spending growth is reshaping financial expectations for the top AI developers. Anthropic's potential valuation recently dropped to between $1.2 trillion and $1.3 trillion, down from an estimated $2 trillion earlier in October, and credit agencies are expected to issue high junk ratings for the company. OpenAI also lowered its proposed valuation target to $1.25 trillion from $1.4 trillion while reducing its latest fundraising goal by $10 billion.
Although spending showed a brief re-acceleration with an 8.9% increase between late September and late October, strong price competition and increased customer efficiency suggest the period of hypergrowth for model provider revenues may be slowing down.
Comments 0
No comments yet. Start the conversation.