Business

Ramp Index Shows AI Costs Fall as Anthropic and OpenAI Clash

A new report from Ramp reveals that corporate AI spending is dropping even as usage reaches record highs, driven by aggressive price competition between Anthropic and OpenAI.

The Decoder4 days agoBusiness
Image: The Decoder

American businesses are ramping up their artificial intelligence usage while simultaneously cutting their overall expenditures on the technology. According to the latest Ramp AI Index, which tracks transaction data from more than 70,000 U.S. companies, corporate spending on AI has steadily declined since peaking in July. Meanwhile, actual usage has surged by approximately 50 percent from that summer peak, reaching an all-time high at the end of September.

Ara Kharazian, an economist at Ramp, attributes this divergence directly to intense market competition between the industry's two dominant players, OpenAI and Anthropic. To capture market share, both companies have introduced significant price cuts on their flagship models while rolling out cheaper, more efficient standard and lite alternatives. This aggressive pricing strategy has allowed enterprise customers to run far more queries for a fraction of their previous costs.

The index, which specifically tracks API spending and skews toward larger AI customers, highlights a highly concentrated market. During the final week of September, Anthropic captured 51 percent of the tracked token spending, while OpenAI secured 44.5 percent. Together, the two rivals control the vast majority of corporate API budgets, leaving open-source models with less than five percent of total business spending.

For enterprise practitioners, these findings signal a highly favorable environment for scaling AI deployments. The prolonged decline in token pricing means development teams can build and run complex agentic workflows or high-volume data processing pipelines with significantly reduced financial risk. As long as the rivalry between Anthropic and OpenAI persists, businesses can expect to extract increasing computational value from their AI budgets.

This is our own summary of reporting by The Decoder

More in Business