What Are Companies Getting for All That AI Spending?
TLDR
The Linux Foundation just launched the Tokenomics Foundation to standardize how AI companies report costs and energy use. Right now nobody agrees on what an AI model actually costs to run, which makes comparing them nearly impossible.
Ask five different AI companies how much energy their models burn or what running them actually costs, and you will get five different answers, measured five different ways, if you get an answer at all. That is the mess the Linux Foundation is trying to clean up with a new group called the Tokenomics Foundation, announced this week.
The problem sounds almost embarrassingly basic for an industry that raised hundreds of billions of dollars on the promise of transforming everything. With thousands of models now available, from tiny open-weight systems to massive frontier models, there is no shared standard for disclosing what any of them actually cost to operate, or what benefits customers should expect in return. Companies pick their own metrics, their own benchmarks, and their own definitions of efficiency, which makes side-by-side comparison close to meaningless.
The Tokenomics Foundation wants to fix that by building common parameters that AI providers can use to report costs and energy requirements consistently. Think of it as an attempt to do for AI what nutrition labels did for food, or what MPG ratings did for cars, giving buyers a baseline they can actually trust instead of marketing claims dressed up as data.
This matters because enterprise AI spending has become a black box even for the people signing the checks. Companies are pouring money into deployments without solid data on whether a given model choice is efficient, wasteful, or somewhere in between. And with data center energy demand already straining power grids in places like Virginia and Ireland, having no agreed way to measure AI's energy footprint is not just an inconvenience. It is a policy problem waiting to get worse.
My take
Standardized disclosure is long overdue, and it is a little wild that an industry moving this fast has spent years without agreeing on basic units of measurement. Expect providers to resist any standard that makes their models look less efficient than a competitor's, because right now the fog benefits whoever spends the most on marketing.
Read more about this at: TLDR
