# What is AI tokenomics?

> AI tokenomics is the practice of managing what an organization spends on AI tokens: what they cost, who spends them, what the spend returns, and how it is held to a budget. The public definitions are about the token; Asato applies the term to the whole AI bill, seats and credits included. Asato works on the cost side: one ledger, and a proposed owner, a budget and a limit for each line.

Canonical page: https://www.asato.ai/ai-tokenomics

Headline on the page: What is AI tokenomics?

Line under the headline: Knowing what AI tokens cost and who spends them, and holding the bill to a budget.

The term is the market's, not Asato's. IDC writes of token economics, or tokenomics, and in August 2026 the Linux Foundation launched the Tokenomics Foundation and defined AI tokenomics. Both definitions are given below. Both are about the token. Taking in seats and credits as well is this page's own step, and Asato's, and the page says so.

## The definition

- **In one sentence.** AI tokenomics is the practice of managing what an organization spends on AI tokens: what they cost, who spends them, what the spend returns, and how it is held to a budget.
- **What it covers.** The token, the unit AI is metered in. Most AI bills also carry seats and credits, so Asato manages all three together.
- **Who owns it.** No one team by default. It sits between IT, finance and engineering, and FinOps teams are taking it on.

On who owns it:

- IDC notes that finance, IT and the vendor each have a different answer for who owns the AI bill. Source: [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/).
- 98% of the respondents to the State of FinOps 2026 survey now manage AI spend, up from 63% in 2025 and 31% in 2024. Source: [FinOps Foundation, State of FinOps 2026](https://data.finops.org/).

## How it is defined

Page line: A new term. Two bodies have defined it.

- **The Linux Foundation.** It launched the Tokenomics Foundation in August 2026, and calls AI tokenomics an emerging practice: managing how AI is produced, how it is consumed and what value it returns. Source: [Linux Foundation, launch of the Tokenomics Foundation, August 4, 2026](https://www.linuxfoundation.org/press/linux-foundation-launches-the-tokenomics-foundation-to-define-the-economics-and-roi-of-ai-value).
- **IDC.** It calls tokenomics a financial discipline: measuring, benchmarking and governing what an enterprise pays for each unit of AI inference, the token. Source: [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/).

Both definitions are about the token. Most AI bills also carry seats and credits, so this page takes in all three.

## Why it matters

Page line: A fast-growing bill, with no budget line of its own.

- **24×** is how far Goldman Sachs forecasts token consumption will grow by 2030, the Linux Foundation reports. Source: [Linux Foundation, launch of the Tokenomics Foundation, August 4, 2026](https://www.linuxfoundation.org/press/linux-foundation-launches-the-tokenomics-foundation-to-define-the-economics-and-roi-of-ai-value).
- **1 in 5** CEOs have created a separate budget line for AI, IDC reports. Source: [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/).
- **98%** of the FinOps practitioners surveyed now manage AI spend, up from 31% two years before. Source: [FinOps Foundation, State of FinOps 2026](https://data.finops.org/).

The first figure is a forecast, not a measurement. The third is from a survey of FinOps practitioners, who are the people most likely to have taken AI spend on.

## Seats, credits, tokens: how AI is billed

Page line: Three ways AI is billed, often on one invoice.

1. **Seats.** A price for each person, each month. The count is known on day one.
2. **Credits.** An allowance drawn down by use. Each vendor counts in a unit of its own.
3. **Tokens.** The unit a model reads and writes, priced by the million. Output costs three to five times more than input.

The price of output against input is from [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/).

Many vendors bill more than one way at once:

- Microsoft 365 Copilot bills a seat, and bills agents in Copilot Credits.
- Claude Enterprise is priced per seat with usage on top, drawn from a pool the organization shares.
- GitHub Copilot bills a seat, and meters usage in GitHub AI Credits, consumed by tokens.
- Cursor bills a seat with included usage, and bills on-demand usage beyond it in arrears.
- The OpenAI API and Amazon Bedrock bill by the token, with no seat.

The visual page gives three of these in one paragraph and links two vendors' own pages: [Microsoft Learn, Copilot Credits](https://learn.microsoft.com/en-us/microsoft-copilot-studio/requirements-messages-management), [GitHub, GitHub Copilot is moving to usage-based billing](https://github.blog/news-insights/company-news/github-copilot-is-moving-to-usage-based-billing/).
Every vendor's own page is linked from the Control AI spend page: https://www.asato.ai/control-ai-spend.md

## An owner, a budget, a limit

Page line: What each line of an AI bill needs.

1. **An owner.** A named person and a cost center for each seat, key and agent. A provider's invoice shows a key or a project, not a team.
2. **A budget.** Set for each team and each month, and forecast while the month is open, not read off the bill.
3. **A limit.** The budget, written into a control that can hold it: the vendor's own spend limit, or your gateway.

What the sources say:

- The FinOps Foundation's working group paper recommends that each key has a named owner, a cost center and an approved use. Source: [FinOps Foundation working group paper, Tokenomics: Managing AI Value in SaaS Model Token Costs](https://www.finops.org/wg/token-economics-saas/).
- IDC reports that one in five CEOs have created a separate budget line for AI. Source: [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/).

No source is cited here for the limit: it is the step Asato adds.

## How it differs from the disciplines beside it

| Discipline | What it is | How AI tokenomics relates | Definition from |
|---|---|---|---|
| FinOps | A practice for getting the most business value from technology spend, begun in cloud. | The nearest practice. 98% of the FinOps practitioners surveyed now manage AI spend. | [FinOps Foundation, What is FinOps?](https://www.finops.org/introduction/what-is-finops/) |
| IT asset management (ITAM) | Managing, controlling and protecting IT assets through every stage of their life. | Holds the seats, contracts and owners that a token bill needs. | [ISO/IEC 19770-1:2017, IT asset management systems](https://committee.iso.org/sites/jtc1sc7/home/projects/flagship-standards/isoiec-19770-12017.html) |
| An AI gateway | An intermediary between applications and AI models: one point for security, governance and observability. | One place a limit can be held, for the traffic routed through it. | [Gartner, market definition of AI gateways](https://www.gartner.com/reviews/market/ai-gateways) |
| Technology Business Management (TBM) | A discipline whose taxonomy classifies technology costs, resources and services. | Says where the money goes. Tokenomics manages one fast-moving line of it. | [TBM Council, TBM Taxonomy](https://www.tbmcouncil.org/taxonomy/) |

The figure in the FinOps row is from [FinOps Foundation, State of FinOps 2026](https://data.finops.org/): the survey's respondents are FinOps practitioners.

Detailed comparison: [Asato vs AI cost tools](https://www.asato.ai/asato-vs-ai-cost-tools.md).

## What Asato does about it

Page heading: What Asato does about it. It works on the cost side: the ledger, the owner, the budget and the limit.

The definition above includes what the spend returns. Asato does not measure the business outcome of an AI workload: it works on the cost side.

1. **Link: One ledger.** Asato reads each vendor's seats, credits and tokens, read-only by default, and ties each seat, key and agent to a person, a cost center and a contract.
2. **Think: An owner and a forecast.** Each line is priced, given a proposed owner, and forecast against its budget while the month is open.
3. **Execute: A limit, once approved.** A person you name approves. Then Asato writes the limit into the vendor's own controls, where the vendor allows it, and reads it back.

Asato writes limits to Claude Enterprise, Cursor, GitHub Copilot and AWS Budgets, and to LiteLLM and Azure API Management gateways. For OpenAI API projects, Anthropic API workspaces, Microsoft 365 Copilot, Azure OpenAI and Vertex AI, the owner gets an alert with the step to take. Asato is never in the path of your AI traffic, and never collects prompts, responses or files. The vendor or the gateway does the blocking, and Asato checks that an owner's step worked.
The solution page has the detail, vendor by vendor: https://www.asato.ai/control-ai-spend.md

## Fund AI from waste

Page line: What nobody uses can pay for the tokens you need.

Many organizations report being asked to fund AI through optimization savings, the FinOps Foundation finds. Source: [FinOps Foundation, State of FinOps 2026](https://data.finops.org/).

Where the money can come from:

- **Software.** Licenses and seats nobody uses, and applications that overlap. See: [Right-size software](https://www.asato.ai/right-size-software.md).
- **Hardware.** Support and maintenance on hardware that is not running. See: [Stop paying for idle hardware](https://www.asato.ai/idle-hardware.md).
- **Contracts.** Renewals signed before anyone checked what was still used. See: [Stay ahead of contracts](https://www.asato.ai/stay-ahead-of-contracts.md).

Label on the page, above the two figures: Results in software, not in AI spend.
The results below are in software, not in AI spend. Asato has published no customer result on AI spend. Found, Approved and Saved are the only three states of a saving, and are never added together.

- **Found: About 1,200** unapproved applications in use at CDW, found in a four-week engagement. Case study: https://www.asato.ai/case-study-cdw-saas-visibility.md
- **Found: About 15%** of productivity-suite spend at CDW sat with inactive or invalid accounts.

Asato's whitepaper on the subject: [The AI Funding Paradox](https://www.asato.ai/whitepaper-ai-funding-paradox.md)

## How a customer starts

Section heading on the visual page: "Where to start." The line under it: "One business unit. Your Asset Clarity Report takes one week, with read-only access." The action is the button "Get your Asset Clarity Report".
This page says nothing about what Asato charges.

To start, request a demo at https://www.asato.ai/talk-to-sales.

## Questions

The page shows at most five questions, the ones that belong to it. Any others are kept here for agents.

### What is AI tokenomics?

The economics of AI tokens. AI tokenomics is the practice of managing what an organization spends on AI tokens: what they cost, who spends them, what the spend returns, and how it is held to a budget. The public definitions are about the token. Most AI bills also carry seats and credits, so Asato manages all three together.

### Is "AI tokenomics" a standard term?

Not yet. The Linux Foundation launched the Tokenomics Foundation in August 2026 to establish open standards for the economics of AI, and definitions still differ from one source to the next. The word is also used in crypto, for the economics of a digital coin. This page is about the tokens AI models bill in.

### Why is an AI bill hard to predict?

It follows use, not seats. A seat count is known on day one. Token use is not: it follows what people and agents do. IDC finds that output tokens cost three to five times more than input tokens, and Gartner predicts that inference cost per agentic workflow will rise more than fivefold through 2028.

### Is AI tokenomics the same as FinOps?

No, but they are close. FinOps is a practice for managing technology spend. It began in cloud, and AI spend is now part of its everyday scope. AI tokenomics is narrower. It deals with the unit AI is billed in, and Asato applies it to the seats and credits that are often bought straight from an AI vendor, outside the cloud bill.

### How does Asato control token spend?

An owner, a budget and a limit. Asato reads each vendor's seats, credits and tokens into one ledger, proposes an owner for each seat, key and agent, and forecasts each line against its budget. Once a named person approves, it writes the limit into the vendor's own controls, where the vendor allows it.

### What is a token?

The unit a model bills in. A token is a small piece of text, about a word or part of one, that an AI model reads or writes. Model providers count the tokens that go in and the tokens that come out, and price each by the million.

### Does Asato measure the return on AI spend?

No. Asato covers the cost side: the ledger, the owner, the budget and the limit, and it counts each saving as Found, Approved or Saved. It does not measure the business outcome of an AI workload.

### How does an organization start?

With one business unit. Asato needs read-only admin access on two or three AI vendors, the directory, the contract rates and a budget owner for each team. In one week you hold your Asset Clarity Report.

### Who approves a change?

A person you name. Nothing changes until a named person approves it. A person you name approves each fix, or sets a policy that governs fixes of that kind. Each step is logged: who approved it, what changed, where and when.

## Sources

Every outside fact on this page comes from an analyst firm or a nonprofit body. Microsoft and GitHub are cited only for how they themselves bill. No company that sells a tool for managing AI cost is cited.

- [Linux Foundation, launch of the Tokenomics Foundation, August 4, 2026](https://www.linuxfoundation.org/press/linux-foundation-launches-the-tokenomics-foundation-to-define-the-economics-and-roi-of-ai-value)
- [IDC, The Token Is the New Budget Line](https://www.idc.com/resource-center/blog/the-token-is-the-new-budget-line/)
- [FinOps Foundation, State of FinOps 2026](https://data.finops.org/)
- [Microsoft Learn, Copilot Credits](https://learn.microsoft.com/en-us/microsoft-copilot-studio/requirements-messages-management)
- [GitHub, GitHub Copilot is moving to usage-based billing](https://github.blog/news-insights/company-news/github-copilot-is-moving-to-usage-based-billing/)
- [FinOps Foundation working group paper, Tokenomics: Managing AI Value in SaaS Model Token Costs](https://www.finops.org/wg/token-economics-saas/)
- [FinOps Foundation, What is FinOps?](https://www.finops.org/introduction/what-is-finops/)
- [ISO/IEC 19770-1:2017, IT asset management systems](https://committee.iso.org/sites/jtc1sc7/home/projects/flagship-standards/isoiec-19770-12017.html)
- [Gartner, market definition of AI gateways](https://www.gartner.com/reviews/market/ai-gateways)
- [TBM Council, TBM Taxonomy](https://www.tbmcouncil.org/taxonomy/)
- [Gartner, press release of August 17, 2026, on inference cost per agentic workflow](https://www.gartner.com/en/newsroom/press-releases/2026-08-17-gartner-predicts-ai-inference-costs-per-agentic-workflow-will-increase-more-than-fivefold-through-2028)

## Related pages

- [Control AI spend](https://www.asato.ai/control-ai-spend.md)
- [Asato vs AI cost tools](https://www.asato.ai/asato-vs-ai-cost-tools.md)
- [What is IT asset optimization?](https://www.asato.ai/it-asset-optimization.md)
- [How Asato compares](https://www.asato.ai/compare.md)
- [Platform](https://www.asato.ai/platform.md)
- [Asato home](https://www.asato.ai/index.md)
- [Site index for agents](https://www.asato.ai/llms.txt)

Last updated: 2026-10-12