Nearly a third of corporate leaders report difficulty understanding and controlling operating costs when implementing business AI at scale, according to a survey from KPMG.

In recent months, Anthropic, OpenAI, and GitHub have shifted some services away from flat-rate subscriptions toward usage-based billing.

"As usage-based pricing models become more common, many organizations are still building the capabilities required to forecast, monitor, and manage AI spending effectively," KPMG said.

The survey of 2,145 senior leaders across 20 countries found that 29 percent struggle to understand their operating costs as they scale their enterprise AI deployments.

A third of senior corporate leaders also identified limited understanding of AI costs and economics as a challenge to deploying AI agents.

Businesses are rethinking their AI plans in the face of changing cost structures and rising fees. The research also found nearly half of organizations have rephased AI deployments when costs have outweighed the expected value. Lower-cost, high-fidelity models are the fastest-growing influence on AI strategy, up 7 percentage points from Q1.

"These actions do not signal reduced confidence in AI. Rather, they suggest a growing willingness to evaluate where AI creates meaningful value and where it does not. Organizations appear increasingly focused on concentrating investment where expected returns are strongest," the report said.

Amazon plans capital expenditure of around $200 billion this year, largely to provide capacity for AI in its AWS datacenters, an increase of 50 percent on a year earlier. Microsoft's total capex is expected to reach $190 billion, up 61 percent from the previous year.

Both companies are now investing significantly in forward-deployed engineering to help customers develop AI applications that will generate demand for the capacity being built.

Amazon has announced a $1 billion investment in an AWS Forward Deployed Engineering organization help customers adopt AI agents and reduce timelines for deployment.

Microsoft is providing $2.5 billion in funding for a new operating entity called Microsoft Frontier Company, "enabling customers to amplify their IQ with AI while refining their differentiated value in the markets that they serve."

In the KPMG report, challenges remain around AI governance: the question of who takes responsibility for decisions made by statistical models prone to erroneous outputs – or "hallucinate," as tech vendors would prefer.

KPMG said executive accountability is important, but "governance ultimately succeeds or fails through day-to-day operating practices."

"Organizations need clear rules for when employees can intervene, who owns AI-related costs, how AI outputs are reviewed and what happens when systems fail. While most organizations report having at least some governance mechanisms in place, relatively few describe these practices as fully embedded," the report said.

Perhaps the tech consultancy and services giant speaks from experience. Last month, research outfit GPTZero claimed a forensic review of KPMG's October 2025 report, "Total Experience: Redefining Excellence in the Age of Agentic AI," found that only five of its 45 citations pointed accurately to the cited source. The rest contained errors ranging from misleading or invented details to references that were too vague to verify.

KPMG later removed the report from some of its websites and issued a statement.

"KPMG International takes the accuracy and integrity of its published content seriously. The report has been removed and we are reviewing the circumstances surrounding its publication. We expect all our people to follow our guidelines on the responsible use of AI, including human oversight to validate content and verify independent sources," a spokesperson said. ®