AI Bills Baffle the C-Suite as Shift to Usage-Based Pricing Challenges Enterprise Cost Management
Key Takeaways
- ▸29% of enterprise leaders struggle to understand and control AI operating costs at scale, revealing an infrastructure gap between adoption and cost management maturity
- ▸Shift from flat-rate to usage-based pricing by Anthropic, OpenAI, and GitHub is driving urgency for better cost forecasting, monitoring, and governance capabilities
- ▸Nearly 50% of organizations have rephased or delayed AI deployments when costs exceeded expected value, signaling a pivot from AI-hype-driven adoption to ROI-focused evaluation
Summary
A new KPMG survey of 2,145 senior executives across 20 countries reveals a significant gap between corporate AI adoption and cost management capabilities. Nearly 29% of executives report difficulty understanding and controlling operating costs when deploying business AI at scale, a problem exacerbated by the recent shift by major AI providers like Anthropic, OpenAI, and GitHub from flat-rate subscriptions to usage-based billing models. Organizations struggling with cost forecasting and governance are increasingly reconsidering their AI deployment plans, with nearly half having rephased projects when costs exceeded expected value.
The shift to usage-based pricing reflects market maturation and a move toward more granular cost accountability. However, KPMG found that most organizations still lack the infrastructure to forecast, monitor, and manage AI spending effectively. A third of executives also identified limited AI cost literacy as a specific barrier to deploying AI agents. In response, cloud giants Amazon and Microsoft are doubling down on forward-deployed engineering teams to help customers navigate new cost structures—Amazon pledging $1 billion for its AWS Forward Deployed Engineering organization and Microsoft committing $2.5 billion for a new Microsoft Frontier Company focused on accelerating customer AI adoption.
- AI governance remains nascent—most organizations lack fully embedded practices for cost ownership, output review, and failure accountability despite recognizing its importance
- Major cloud providers investing heavily ($1B+ commitments) in forward-deployed engineering teams to help enterprises manage AI costs and accelerate deployment timelines
Editorial Opinion
The KPMG survey captures the AI market at a critical inflection: early-stage enthusiasm is colliding with the hard economics of deployed AI. Usage-based pricing is economically rational and aligns incentives, but it's clear the enterprise market lacks the governance, forecasting, and cost-optimization maturity to operate effectively under it. The massive capex commitments from Amazon and Microsoft suggest these infrastructure providers expect sustained AI ROI—but survey data shows that confidence hasn't yet trickled down to their customers' CFOs.
