Inference costs for AI agents may increase more than fivefold by 2028, even as the price per token continues to fall. The paradox lies in how companies are deploying these models: as agents tackle increasingly complex tasks, they demand substantially more processing power, driving total workflow costs upward.
Token prices dropping doesn't automatically mean lower expenses. Enterprises running sophisticated AI agents on demanding workloads face a math problem—cheaper tokens multiplied by vastly higher token consumption can exceed previous spending. To preserve return on investment, organizations must adopt smarter approaches: optimize agent orchestration and route each task to the most cost-effective model capable of handling it, rather than defaulting to the largest or most capable model for every job.