Understand this story
A survey published by Harness.io shows that 67% of the enterprises surveyed spend at least $250,000 per month on AI. The survey, covering about 700 global enterprises, also found that 73% of AI projects exceed their budgets because AI workload costs are difficult to predict. The survey also shows that 57% of the enterprises surveyed encourage employees to use AI without restrictions, and Harness.io estimates that this wastes 26% of the monthly AI budget. The related report discusses these results in the context of the difficulty traditional cloud FinOps has in controlling AI costs, noting that token-based billing, multi-model environments, rapid model iteration, and the acceleration of code generation and agentic workflows are the main reasons. Currently, the survey results are reported by a single source, with no subsequent developments yet.
AI-generated from public reports
Latest development · Report
AI FinOps ①: Why Traditional FinOps Can't Control AI CostsDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 20
The Elec
AI FinOps ①: Why Traditional FinOps Can't Control AI Costs
A Harness.io survey of about 700 global enterprises shows that 67% of enterprises spend at least $250,000 per month on AI, and 73% of AI projects exceed budget because AI workload costs are hard to predict. The article points out four major reasons why traditional cloud FinOps fails in the AI environment: token-based billing, multi-model environments, rapid model iteration, and the acceleration of code generation and agentic workflows. The survey also shows that 57% of surveyed enterprises encourage employees to use AI without restrictions, and Harness.io estimates that 26% of monthly AI budgets are wasted as a result.