The Linux Foundation recently announced its intent to launch the Tokenomics Foundation, a new organization that will establish open standards for AI economics, in close partnership with the FinOps Foundation.
The move is already backed by Accenture, Booking.com, IBM, JPMorgan Chase, Microsoft, Oracle, and more. The buyer side of the agentic boom is organizing as enterprises realize that agent costs are out of control, and they have no reliable way to measure ROI.
"Token costs and efficiency have become a CEO-level concern, not an engineering footnote," says J.R. Storment, Executive Director of the FinOps Foundation.
Examples of runaway AI spend are already piling up. In April, Uber's CTO Praveen Neppalli Naga disclosed that the company burned through its entire 2026 AI coding budget in just four months. Meanwhile, Amazon employees were gaming internal AI usage leaderboards, optimizing for “tokenmaxxing” rather than business value, and getting rewarded for it.
Gartner predicts that more than 40% of agentic AI projects will be canceled by 2027 due to escalating costs, unclear business value, and inadequate governance.
The Tokenomics Foundation is the industry's initiative not to abandon AI investment but to build the infrastructure that makes it worth defending.
Tokenomics Is About More Than Reducing Costs
The reflex when AI costs spiral is to cap or eliminate AI budgets. But that halts the innovation engineers were hired to drive, and gives them no path to the ROI agentic workflows were supposed to deliver.
Tokenomics applies FinOps discipline to AI, and the Foundation is formalizing the principles to make that practical. It is not about saving money on tokens. It is about knowing whether token spend was worth it.
An agent that burns 50,000 tokens a day answering help-center questions and another that burns the same amount closing a five-figure deal offer very different value to the business. But as Uber COO Andrew Macdonald said on a recent podcast, companies can’t yet tell the difference: “That link is not there yet.”
The potential value of agentic AI is compelling. But few organizations have the systems to measure it, let alone control it.
Token Spend Is Only Part of the Picture
The Tokenomics Foundation's work is essential, and every company operating in AI economics has reason to back it. The standards, language, and benchmarks it produces will be the foundation on which the next decade of AI economic control is built.
But as Storment says, tokens are the most visible and most easily metered layer of AI spend, and only a portion of it. Tokens are the tip of the iceberg. Every agent action also triggers downstream activity, like data enrichment, database queries, and SaaS tools with metered API usage; costs that don’t show up on the token bill. To see the returns AI vendors have been promising, enterprises will need to understand the fully burdened cost below the surface and connect it (tokens and all) back to business outcomes.
Revenium's AI Economic Control System is an attribution layer that connects each token, retry, and downstream API call back to the agent, workflow, or person who triggered it, and to the business outcome it produced.
Instead of capping every engineer's AI spend, Revenium ties it to outcomes in real time and can even throttle a misfiring agent before it burns your budget.
Sign up for Revenium to see the true value of your AI agents and start driving the AI ROI you’ve been waiting for.
