JPMorgan Chase, Accenture and others are teaming up on a new venture to standardize how AI token use is measured

Good morning. A coalition of major corporations aims to establish standardized methods for measuring and managing artificial intelligence (AI) token costs and their connection to business value. The Linux Foundation's Tokenomics Foundation, launching today with 29 initial members, includes firms like Accenture, Booking.com, BNY, Flexera, IBM, JPMorgan Chase, KPMG, Oracle, SAP, and ServiceNow. The group's goal is to unite token users and suppliers in a vendor-neutral platform.

CFO Daily spoke with J.R. Storment, the Tokenomics Foundation's executive director, about how CFOs can develop expertise in AI token management. Storment advises CFOs to work with their teams to gain visibility into consumption by model, workload, team, and project, aiming to align budgets with outcomes rather than flat limits. He emphasizes the importance of partnering with the CTO, CIO, and CAIO to understand current AI activities, planned initiatives, and expected outcomes.

CFO Daily inquired about the market gap that prompted the creation of the standalone Tokenomics Foundation rather than integrating it into the FinOps Foundation. Storment explained that while per-token costs fell significantly from 2023 to 2025, they have since leveled off, and new model token prices are rising, making AI costs the largest and fastest-growing line item in enterprise technology budgets.

The rapid pace of AI change, the increasing number of model options, and higher production costs further emphasize the need for AI value measurement.

With 29 founding members spanning various sectors, including hyperscalers, enterprises, and vendors, Storment discussed how they managed to bring them together. He highlighted the Linux Foundation's history of facilitating vendor-neutral cooperation, focusing primarily on end-user enterprises' needs to solve their challenges collaboratively. The Foundation aims to standardize pre-competitive technology, frameworks, and benchmarks to avoid duplication of efforts and accelerate AI investment.

Regarding the timeline for developing usable frameworks and benchmarks, Storment emphasized the need to balance speed and standardization. Already, lightweight best practices like the Big-T Notation framework have been published, with more drafts anticipated for presentation at the Tokenomicon in Amsterdam in September. The Foundation plans to release new frameworks and value metric definitions nearly every month through the end of the year.

Additionally, a training and education track is being developed to ensure AI literacy among executives, as AI adoption across organizations is accelerating rapidly.