Rising AI bills are reshaping how businesses choose their models, with cost overtaking raw capability as the decisive factor, Reuters reports.
Companies that initially adopted the most advanced — and most expensive — AI models are now downgrading to cheaper alternatives. The pattern is consistent across sectors: enterprises test premium offerings, watch costs climb, and switch to models that deliver 80% of the performance at a fraction of the price.
By Aditya Soni. The shift has implications for the business models of AI labs that charge premium rates for frontier models. If the enterprise market concludes that second-tier models are good enough for most workloads, the addressable market for top-tier pricing shrinks.
The cost sensitivity is amplified by the scale of deployment. A model that costs a few cents per query in testing becomes a material line item when deployed across thousands of employees or millions of customer interactions.
Analysis
This is the commoditisation curve playing out in real time. Frontier models command premium pricing only as long as the gap between them and the next tier is wide enough to justify the cost. That gap is narrowing. Open-weight alternatives from Meta, Mistral, and Chinese labs are compressing the premium window. The AI labs betting on sustained high-margin enterprise revenue — OpenAI at $5/$30 per million tokens for GPT-5.6 Sol — are selling into a market that is already shopping for discounts.
