The Foundation’s draft definition (v0.5.2, 1 September 2026): “AI Tokenomics (i.e., ‘AI Economics’) is the discipline of converting energy and capital into AI capabilities and efficiently consuming that intelligence across the organization, to realize measurable business value.”
It is not the cryptocurrency term. A token here is the unit in which AI output is metered. The Foundation uses AI Economics as a second name for the discipline, which is also why the product is called AI Economics.
At the sourcetokeneconomics.com. The Foundation operates as a Series of LF Projects, LLC, under the Linux Foundation.
Production
How is the token made?
Turning energy and capital into AI capacity: hardware, capacity planning.
Consumption
How is the AI used?
Allocation, forecasting, optimization and cost to serve. The practitioner core, and the part the FinOps Framework calls FinOps for AI.
Value
What did the AI return?
Pricing, monetization and return on investment. The least settled of the three.
MagicOrange is a founding member of the Tokenomics Foundation.
| Foundation concept | Where it runs |
|---|---|
| All AI costs, not only token costs | AI Economics: the hyperscaler, the model API, the vector database and the orchestration layer land in one model beside the data platform and the people. |
| Attribution to team, feature and agent | Token usage, model API, agent workload and GPU as shipped standard objects. Agents are economic objects with an owner, a purpose and a measurable outcome. |
| Cost per outcome | AI unit economics: cost per inference, per successful outcome and per AI-assisted transaction, loaded with tokens, platform, data and labor. |
| Budgets that are not blanket caps | Token budgets by team, model and use case, reported against continuously. |
| FOCUS, extended for tokens | Cloud-borne AI cost enters through FOCUS today. The model takes token-level fields as the specification defines them. |
| The value domain | Cost per outcome set against the value of the outcome, by program, in Analytics. |
MagicOrange names, measures and governs AI cost. It is not a token meter, an AI gateway, a router or an inference optimizer. It is the enterprise layer above them: it attributes, allocates, budgets and recharges what the stack produces, and connects it to what the spend returned. Decisions to throttle or stop a workload stay with your engineering team.
The Consumption domain in the FinOps Framework’s terms.
Which is being extended for tokens.
The parent body.
Our framework.
