IMPLICATOR
A new KPMG survey reported by The Wall Street Journal found that only 26% of companies fully track AI costs. It put 50% at partial visibility and 22% at little or none until billing, as vendors meter enterprise AI through tokens and cloud compute. The findings add to KPMG, Deloitte and Linux Foundation reports that token-based AI costs are moving from experiments toward managed technology spend.
Key Takeaways
- KPMG survey reported by WSJ found only 26% of companies fully track AI costs.
- Deloitte cited a health care case where token growth added more than $6 million in annualized unplanned costs.
- Cloudflare added dollar-based spend limits for AI Gateway and closed-beta identity budgets.
- The Linux Foundation says the Tokenomics Foundation will set open standards for AI cost management.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Survey puts the bill inside finance
The Journal reported that the KPMG survey is not yet public, so the 26% figure remains a single-source result. Steve Chase, KPMG’s global head of AI, told the paper that some clients have burned through annual token and cloud budgets in months, and that one client’s token usage rose sixfold. KPMG’s earlier paper, The first token’s free, listed token charges as AI run costs and urged total-cost visibility before programs scale. The same cost gap is the use case for gateways and observability tools, including Cloudflare AI Gateway, that attribute usage by model or team.
Token meters change the budget
Deloitte wrote in April that tokens, not licenses or head count, are becoming the unit of cost for some AI programs. In one health care example cited by Deloitte, token usage grew 8% to 10% per month over six months and added more than $6 million in annualized unplanned costs before finance could trace it. Deloitte also found many companies generate more than 10 billion tokens a month, with the share expecting more than 100 billion projected to triple from 2025 to 2028.
Companies add controls
Several companies in the Journal report have changed AI controls. According to the paper, Life360 finance chief Russell Burke wants a real-time monitor for token spending, and Affirm finance chief Rob O’Hare put token spending into annual budgeting after agent-written code lifted consumption in its March quarter. Corning finance chief Ed Schlesinger limited employee access to AI tools while keeping a smaller set of major projects funded.
Track the AI bills shaping strategy
Strategic AI news from San Francisco. No hype, no "AI will change everything" throat clearing. Just what moved, who won, and why it matters. Daily at 6am PST.
No spam. Unsubscribe anytime.
Cloudflare announced spend limits for AI Gateway on June 5, with budgets set in dollars and scoped by model, provider, user, team or application. By default, over-budget requests are blocked; customers can also configure routing rules to fall back to a cheaper model. Cloudflare also announced a closed beta for identity-driven budgets and policies, using Cloudflare Access to attach employee or service-token identity to AI Gateway requests.
Know someone who'd find this useful? ✉️ Email it to a friend in one click, or they can subscribe free here.
Standards effort starts in San Diego
The Linux Foundation announced on June 3 that it intends to launch the Tokenomics Foundation with initial support expressed by Accenture, Booking.com, Flexera, Google Cloud, IBM, JPMorganChase, KPMG, Microsoft, Oracle, Salesforce, SAP and ServiceNow. The release cited Goldman Sachs research projecting global token usage will rise 24 times from 2026 to 2030, reaching 120 quadrillion tokens a month.
The foundation is scheduled to outline its technical roadmap and working groups at FinOps X in San Diego. The event’s own site lists June 8 to 11. The Linux Foundation said no neutral home existed for token-efficiency standards before this effort. Until those standards arrive, finance teams reconcile model logs, cloud invoices and vendor dashboards by hand against budgets that often predate broad agent use.
Frequently Asked Questions
What did KPMG’s AI-cost survey find?
The survey, reported by The Wall Street Journal, found that 26% of companies fully track AI costs. It put 50% at partial visibility and 22% at little or no visibility until billing.
Why are AI costs harder for CFOs to track?
Enterprise AI bills often depend on tokens, cloud compute and model choice rather than fixed licenses. More employees and agents can increase usage before finance teams see the invoice.
Which companies are adding AI cost controls?
The Journal cited Life360, Affirm and Corning as examples. Life360 wants real-time token monitoring, Affirm added token spending to budgeting and Corning limited tool access.
What did Cloudflare add to AI Gateway?
Cloudflare announced spend limits that set budgets in dollars by model, provider, user, team or application. It also announced a closed beta for identity-driven budgets through Cloudflare Access.
What is the Tokenomics Foundation?
The Linux Foundation announced the Tokenomics Foundation as a standards effort for AI token economics. Its initial supporters include Google Cloud, IBM, Microsoft, Oracle, Salesforce and KPMG.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Marcus Schuler
Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: editor@implicator.ai
