![ApptioCareers-Employee[1]](data:image/svg+xml;base64,PHN2ZyB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciIHZpZXdCb3g9IjAgMCAyMjUgMjc1IiB3aWR0aD0iMjI1IiBoZWlnaHQ9IjI3NSIgZGF0YS11PSJodHRwcyUzQSUyRiUyRnd3dy5hcHB0aW8uY29tJTJGd3AtY29udGVudCUyRnVwbG9hZHMlMkZlbGVtZW50b3IlMkZ0aHVtYnMlMkZBcHB0aW9DYXJlZXJzLUVtcGxveWVlMS1xZzVuenB5MHdoaWJtbzZvdmJqNGM3MHh0aTUzdmo1OHM0endwajh0dHMucG5nIiBkYXRhLXc9IjIyNSIgZGF0YS1oPSIyNzUiIGRhdGEtYmlwPSIiPjwvc3ZnPg== "ApptioCareers-Employee[1]")
Webinar: The New Economics of AI: How Leaders Can Optimize Spend and ROI
What is tokenomics?
Tokenomics, sometimes called token economics, is the emerging discipline of managing the cost, efficiency, and business value of AI token consumption. As organizations move generative and agentic AI from pilot to production, tokens have become a new, fast-growing line item on the enterprise AI budget, one with pricing and consumption patterns that don’t behave like traditional cloud or software spend.
The term has gained industry-wide attention following the June 2026 announcement of the Tokenomics Foundation, a Linux Foundation initiative developing open standards for AI cost management in close partnership with the FinOps Foundation. The Foundation frames tokens as the atomic unit of the AI economy: simultaneously the unit of cognition a model produces, the unit of compute a data center serves, the unit of price a lab charges, and the unit of value an enterprise extracts.2
Apptio, an IBM Company, is directly connected to this effort. Bill Lobig, VP of IBM Apptio, is among the named supporters in the Foundation’s launch, stating that enterprises need open, vendor-neutral standards to compare AI cost and efficiency across models and platforms rather than relying on any single provider’s benchmarks.1
AI cost management, at its core, is FinOps applied to AI: the same discipline of visibility, allocation, and accountability that FinOps brought to cloud spend, extended to a new and less standardized cost layer.
Tokenomics started as a crypto term. AI tokenomics is a different discipline.
If you’ve encountered “tokenomics” before, it was likely in the context of cryptocurrency, where the term describes how a token’s supply, distribution, and monetary policy affect its value. Concepts like circulating supply, total supply, and market cap are core to that discussion, and they matter for evaluating a digital asset as an investment.
AI tokenomics borrows the word but not the framework. There’s no supply and demand curve for an AI token, no market cap, no monetary policy to analyze. Instead, an AI token is a unit of consumption, closer to a kilowatt-hour than a cryptocurrency: something you use, pay for, and need to manage the cost of at scale. The Tokenomics Foundation’s own definition reflects this split, describing AI tokenomics as converting energy and capital into AI, then consuming it efficiently to drive business outcomes, a FinOps problem, not an investment one.2
Why tokenomics matters now
A few dynamics are pushing this from an engineering footnote to a CFO-level concern:
AI spending overall is entering trillion-dollar territory.
“Worldwide spending on AI is forecast to total $2.59 trillion in 2026, a 47% increase year-over-year, according to Gartner®, Inc. a business and technology insights company.”3
Token usage is scaling fast.
Industry research cited by the Tokenomics Foundation projects global token usage multiplying roughly 24x between 2026 and 2030.2Per-token pricing is no longer reliably falling.
Costs dropped sharply from 2023–2025 but have leveled off, with some newer frontier model pricing rising.2
AI infrastructure investment is at unprecedented scale.
The AI inference market alone is projected to grow from roughly $106 billion in 2025 to $255 billion by 2030, part of a broader wave of AI infrastructure investment that industry analysts project will exceed $1 trillion through 20272
Enterprise AI budgets are under new scrutiny.
As usage-based pricing makes AI costs harder to predict., Gartner reports, “Enterprise AI budgets are coming under greater scrutiny, with increased focus on usage efficiency, cost control and measurable outcomes.” “This is giving an edge to providers that embed evaluation, cost transparency and usage tracking into customer workflows, making it easier to manage and optimize AI use.”4
Token spend is harder to govern than cloud spend.
Input vs. output tokens, cached vs. non-cached pricing, and inconsistent structures across vendors mean the FinOps playbook built for cloud doesn’t map cleanly onto enterprise AI.
The core components of token cost
Input and output tokens.
Most model providers price these differently, typically charging more for output tokens than input tokens, so the ratio between what you send a model and what it returns materially affects cost. Forecasting spend accurately means tracking that ratio, not just total token volume.
Caching.
Reusing context across requests, rather than resending the same input tokens repeatedly, can reduce costs significantly, but only if usage patterns are tracked and architected for it. Caching is one of the highest-leverage optimization levers available and one of the easiest to miss without visibility into consumption patterns.
Model routing and frontier model selection.
Cost per token varies significantly across model tiers, from smaller, cheaper models to top-tier frontier models. Model routing, directing each request to the least expensive model capable of handling it well, is an emerging technique for controlling spend without sacrificing output quality.
Observability.
Just as FinOps depends on visibility into cloud spend, tokenomics depends on observability into token consumption: which teams, applications, and models are driving usage, and how that usage trends over time. Without it, cost per token is a number with no context.
Unit economics.
The end goal isn’t minimizing token spend for its own sake, it’s understanding the unit economics of AI: what it actually costs to deliver a given outcome (a completed task, a resolved ticket, a generated report), so that spend can be evaluated against the value it produces.
Allocation and attribution.
Mapping token spend back to business units, cost centers, products, or use cases, the same allocation discipline FinOps applies to cloud, extended to a new and less standardized cost layer.
How Apptio approaches AI token reporting and optimization
IBM Apptio applies its existing IT cost transparency and allocation discipline directly to AI token consumption, through connected solutions:
AI TCO & Usage tracks input and output token consumption as a named, dedicated metric over time, helping organizations spot usage trends before they turn into bill shock and identify underused models or solutions worth consolidating or retiring. It also maps AI spend, including token costs, across cloud, labor, infrastructure, and vendor layers, so cost owners can see the full picture rather than just the token line item in isolation.
AI Investment Management, built on IBM Cloudability, goes a layer deeper on attribution, ingesting token, GPU, and API usage data (including directly from providers like OpenAI) through FOCUS and custom connectors. That lets organizations break down AI costs by specific model and usage type, from frontier chat models to embedding models to fine-tuning operations, and track token consumption patterns to see which operations are actually driving cost. It also supports model-selection decisions based on cost-per-token and usage patterns, so teams can weigh cost against performance across models rather than defaulting to the most expensive option.
Underlying both, AI-Powered Apptio’s AI Mapping capability uses generative AI to automatically map and normalize vendor billing and usage data, including token-based pricing, into the FOCUS schema, so token costs from different providers can be compared on common ground rather than reconciled manually.
IBM and the Tokenomics Foundation
IBM is named among the initial supporting organizations of the Tokenomics Foundation, alongside Accenture, Google Cloud, Microsoft, Oracle, Salesforce, SAP, and others. This positions IBM, and by extension Apptio, at the center of an industry-wide effort to standardize how enterprise AI costs are measured, compared, and governed.
Leadership Guide to FinOps for AI
As generative AI costs rise, executives are turning to FinOps practitioners to understand what’s driving costs, how they can be optimized, and which AI investments are delivering real ROI. But many organizations are still scrambling to incorporate AI costs into their FinOps practice.
This guide provides a four-step framework to position your FinOps practice to drive your organization’s AI strategy, ensuring the highest-value AI initiatives are appropriately resourced and enabling smarter, more strategic investment decisions. Read this guide to learn how to:
- Diagnose your organization’s FinOps for AI maturity and align stakeholders on a path forward
- Build a unified view of generative AI costs across all architectures and deployments to enable accurate unit economics and ROI tracking
- Apply targeted optimization levers and surface strategic business outcomes to shift from reactive cost tracking to proactive AI investment decisions
- Linux Foundation, “Linux Foundation Announces the Intent to Launch the Tokenomics Foundation to Establish Open Standards for AI Cost Management,” June 3, 2026.
- Tokenomics Foundation, “What is Tokenomics?” (Draft v0.2, August 2026). Note: this page cites Goldman Sachs (24x token usage growth) and S&P Global Ratings ($1T+ AI infrastructure investment) as its underlying sources; those are currently cited here one level removed via the Foundation’s page.
- Gartner, “Gartner Forecasts Worldwide AI Spending to Grow 47% in 2026,” press release, May 19, 2026. URL: https://www.gartner.com/en/newsroom/press-releases/2026-05-19-gartner-forecasts-worldwide-ai-spending-to-grow-47-percent-in-2026 . GARTNER is a trademark of Gartner, Inc. and/or its affiliates.
- Gartner, “Gartner Forecasts Worldwide AI Platforms and Models Market to Grow 63% in 2026,” press release, July 20, 2026 (quote from Arunasree Cheparthi, Sr Principal Research Analyst). URL: https://www.gartner.com/en/newsroom/press-releases/2026-07-20-gartner-forecasts-worldwide-ai-platforms-and-models-market-to-grow-63-percent-in-2026