Skip to content
AI Economics Hub

Glossary entry · FinOps and TBM

AI Value Data Layer

A draft proposal from the Tokenomics Foundation, published on 5 November 2026, to join AI cost to value in one data store. It would add workload, task and outcome identifiers to OpenTelemetry, and four datasets alongside FOCUS billing data: telemetry, tasks, outcomes and value. A playbook and reference implementation are targeted for June 2027.

Why it matters

No measure that spans cost and value, such as cost per successful outcome, can be calculated from records until the bill, the work and the result share keys. A common format would make the evidence for the first three of this site's five layers comparable across tools and organisations.

Source: the Tokenomics Foundation's draft AI value data layer proposal (Shawn Alpay, 5 November 2026, CC BY 4.0). It is a draft seeking industry input, not an adopted standard. The Foundation describes it as an implementation that supports its framework, in the way FOCUS gives data structure to the FinOps Framework.

The proposal supplies evidence for Layers 1 to 3 of this site's five layers: usage, output quality and productivity value. It does not address Delivery Alignment or Portfolio Strategy, which are decisions made with the data. It leaves two questions open: who is accountable when attribution fails, and how outcome data is governed. See the standards map for where it sits among other standards, and The AI Value Data Layer Is Coming for what executives should do now.

Where this term comes up