Skip to content
THE LINUX FOUNDATION PROJECTS

What is AI Tokenomics?

Don't count tokens. Realize value from the total cost of AI.  

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.

— TOKENOMICS DRAFT DEFINITION V0.5.2, SEPTEMBER 1 2026

The rest of Tokenomics Infrastructure Compute Memory Storage Data Energy Data Centers Software Licenses Outcomes Labor Value Realization Revenue Business Outputs What most focus on Tokens API Costs

Subscribe for AI Tokenomics updates

State of Tokenomics Release 1.0

See it live on September 23 in Amsterdam or Virtually

Reserve your spot

TOKENOMICS FOUNDATION WORKING GROUPS

Tokenomics looks at the entire AI supply chain, from energy to labor to value

Production

Turning energy and capital into AI capabilities and intelligence.

Energy Capital resources AI factories AI capacity creation Capacity planningData CentersHAM

Consumption

Using that intelligence well, and managing the entire bill of materials.

Model Selection ForecastingPrompt Routing Caching Strategy EfficiencyLicensesSoftware FinOps ITAM ITFM

Value

What the output is worth, and how it impacts the business.

Demand Monetization Pricing Products LaborAccountingBusiness ProcessesRevenue

Events

Tokenomics Events Around the World

Projects

What the working groups are building.

Open specifications, benchmarks, and frameworks, built by the organizations doing the work.

View All Projects

Featured Project

The five-layer tokenomics stack

Open a layer to see what it controls

Featured Project

Big-T Notation

Big-O for AI. A shared language for how token consumption grows as usage scales.

T(1) Constant. The model is not called per request — a cache hit, a static lookup. cache and precompute
T(log n) Sublinear. Deterministic code shrinks the input before the model sees it. filter before inference
T(n) Linear. One model call per request — the healthy default. trim per-call overhead
T(n·k) Multiplicative. k model calls per request, and k is usually invisible. compose tool pipelines
T(n·k·a) Agent-multiplicative. An orchestrator spawns sub-agents that spawn tool calls. bound depth, add budgets
T(∞) Unbounded. Loops with no termination condition. hard termination. always.

T(n · k · a) n = requests or input size · k = model calls per request · a = agent depth