AI Cost & Billing Glossary

Plain-English definitions of the terms teams use when metering, billing, and reporting on AI usage — from token pricing to margin analysis.

AI gross margin
AI gross margin is the percentage of AI-feature revenue left after subtracting the direct cost of model inference — vendor API spend, compute, and hosting — and it runs structurally thinner than traditional software margin because inference cost scales with usage rather than staying fixed.
AI monetization model
An AI monetization model is the pricing strategy a company uses to charge for AI-powered products or features — commonly seat-based, usage-based, outcome-based, or a hybrid of these — chosen based on how closely the underlying inference cost structure tracks actual usage.
Blended cost vs. list price
Blended cost is the effective average price actually paid across a mix of usage after volume discounts and negotiated rates are applied, while list price is a vendor's undiscounted published rate — the two diverge whenever any usage qualifies for a discount tier or contract-negotiated rate.
Cost attribution
Cost attribution is the practice of assigning AI and cloud costs to the specific customer, feature, team, or workload that generated them, enabling accurate per-customer margin analysis, internal chargeback, and showback.
Cost per inference
Cost per inference is the total cost of one complete model call or agent action — all input and output tokens across every step, plus any fixed per-request overhead — used to measure unit economics for AI features priced per action rather than per token.
Cost per token
Cost per token is the price paid (or charged) per unit of text processed by a large language model, typically expressed in dollars per million tokens and split between separate input (prompt) and output (completion) rates.
FinOps for AI
FinOps for AI is the practice of applying FinOps discipline — cost visibility, allocation, forecasting, and optimization — to AI and LLM spend, extending traditional cloud FinOps to the token-based, per-inference, multi-vendor cost dynamics unique to AI workloads.
FOCUS spec (FinOps Open Cost and Usage Specification)
The FOCUS spec (FinOps Open Cost and Usage Specification) is an open, vendor-neutral data standard maintained by the FinOps Foundation that normalizes cost and usage data from different cloud and AI providers into one common schema, so spend can be compared and aggregated across vendors.
Margin leakage
Margin leakage is the gradual, often unnoticed erosion of gross margin caused by a gap between what a company is actually billed by its vendors and what it charges its customers — for example, unbilled overage, a stale rate card, or a vendor price change that never gets reflected downstream.
Metered billing
Metered billing is the mechanism that measures and records a customer's consumption of a billable resource — such as tokens, API calls, or compute time — in real time, feeding those measurements into a rate card to calculate what the customer owes at the end of a billing period.
Outcome-based pricing
Outcome-based pricing is a pricing model that charges customers for a completed business result — a resolved support ticket, a qualified lead, a completed task — rather than for tokens, API calls, or seats, so price tracks the value delivered rather than the resources consumed to produce it.
Overage billing
Overage billing is the practice of charging customers for usage that exceeds a plan's included allotment, typically at a per-unit rate specified on the rate card, applied on top of a base subscription fee or committed volume.
Rate card (usage-based billing)
A rate card is the published list of prices a company charges for metered usage — broken out by billable unit, such as per token, per API call, or per seat, and often tiered by volume — that governs how a usage-based invoice is calculated.
Token-based pricing
Token-based pricing is a billing model that charges customers per unit of text (token) processed by an AI model, with separate rates for input and output tokens, mirroring how most LLM vendors themselves price API access.
Usage-based billing
Usage-based billing is a pricing model that charges customers based on their actual consumption of a product or service — such as API calls, tokens, or compute time — rather than a fixed subscription fee, so the invoice amount tracks metered usage directly.
Vendor cost reconciliation
Vendor cost reconciliation is the process of matching a company's own usage records against every AI vendor's billing statements to confirm each request was billed at the expected rate, catching pricing drift, unbilled usage, or vendor-side errors before they distort margin reporting.

Plain-English definitions of the terms teams use when metering, billing, and reporting on AI usage — from token pricing to margin analysis.

AI gross margin

AI gross margin is the percentage of AI-feature revenue left after subtracting the direct cost of model inference — vendor API spend, compute, and hosting — and it runs structurally thinner than traditional software margin because inference cost scales with usage rather than staying fixed.

AI monetization model

An AI monetization model is the pricing strategy a company uses to charge for AI-powered products or features — commonly seat-based, usage-based, outcome-based, or a hybrid of these — chosen based on how closely the underlying inference cost structure tracks actual usage.

Blended cost vs. list price

Blended cost is the effective average price actually paid across a mix of usage after volume discounts and negotiated rates are applied, while list price is a vendor's undiscounted published rate — the two diverge whenever any usage qualifies for a discount tier or contract-negotiated rate.

Cost attribution

Cost attribution is the practice of assigning AI and cloud costs to the specific customer, feature, team, or workload that generated them, enabling accurate per-customer margin analysis, internal chargeback, and showback.

Cost per inference

Cost per inference is the total cost of one complete model call or agent action — all input and output tokens across every step, plus any fixed per-request overhead — used to measure unit economics for AI features priced per action rather than per token.

Cost per token

Cost per token is the price paid (or charged) per unit of text processed by a large language model, typically expressed in dollars per million tokens and split between separate input (prompt) and output (completion) rates.

FinOps for AI

FinOps for AI is the practice of applying FinOps discipline — cost visibility, allocation, forecasting, and optimization — to AI and LLM spend, extending traditional cloud FinOps to the token-based, per-inference, multi-vendor cost dynamics unique to AI workloads.

FOCUS spec (FinOps Open Cost and Usage Specification)

The FOCUS spec (FinOps Open Cost and Usage Specification) is an open, vendor-neutral data standard maintained by the FinOps Foundation that normalizes cost and usage data from different cloud and AI providers into one common schema, so spend can be compared and aggregated across vendors.

Margin leakage

Margin leakage is the gradual, often unnoticed erosion of gross margin caused by a gap between what a company is actually billed by its vendors and what it charges its customers — for example, unbilled overage, a stale rate card, or a vendor price change that never gets reflected downstream.

Metered billing

Metered billing is the mechanism that measures and records a customer's consumption of a billable resource — such as tokens, API calls, or compute time — in real time, feeding those measurements into a rate card to calculate what the customer owes at the end of a billing period.

Outcome-based pricing

Outcome-based pricing is a pricing model that charges customers for a completed business result — a resolved support ticket, a qualified lead, a completed task — rather than for tokens, API calls, or seats, so price tracks the value delivered rather than the resources consumed to produce it.

Overage billing

Overage billing is the practice of charging customers for usage that exceeds a plan's included allotment, typically at a per-unit rate specified on the rate card, applied on top of a base subscription fee or committed volume.

Rate card (usage-based billing)

A rate card is the published list of prices a company charges for metered usage — broken out by billable unit, such as per token, per API call, or per seat, and often tiered by volume — that governs how a usage-based invoice is calculated.

Token-based pricing

Token-based pricing is a billing model that charges customers per unit of text (token) processed by an AI model, with separate rates for input and output tokens, mirroring how most LLM vendors themselves price API access.

Usage-based billing

Usage-based billing is a pricing model that charges customers based on their actual consumption of a product or service — such as API calls, tokens, or compute time — rather than a fixed subscription fee, so the invoice amount tracks metered usage directly.

Vendor cost reconciliation

Vendor cost reconciliation is the process of matching a company's own usage records against every AI vendor's billing statements to confirm each request was billed at the expected rate, catching pricing drift, unbilled usage, or vendor-side errors before they distort margin reporting.