\n<\/p>\n
Across the industry, companies are starting to balk at the price of AI. Uber blew through<\/a> its entire 2026 AI coding budget by April. Microsoft revoked<\/a> its developers\u2019 Claude Code licenses months after enabling them. A Priceline employee told TechCrunch that a routine Cursor contract renewal came back 4-5x more expensive. <\/p>\n Even though per-token prices have fallen, the push for more AI adoption and increasingly autonomous agents have driven token consumption higher and higher.\u00a0Companies that gorged themselves in early 2025 on all-you-can-eat subscriptions are now scrambling to understand where their money is going, pull back spending, and figure out whether they can salvage some ROI from the wreckage of their budgets. <\/p>\n Meanwhile, a market is forming to meet them there. Startups, established vendors, and a new standards body are all racing to give companies the tools and language to track what they spend.<\/p>\n \u201cSix months ago, I would have a conversation with a customer and it would be all about \u2018What can it do? Is it good enough?\u2019\u201d Alexander Embricos, OpenAI\u2019s head of enterprise, told TechCrunch at an event in New York City this week. \u201cOur conversations are never about that now. Now the conversations are about, \u2018hey, we\u2019re spending so much. What visibility do you have? What auditability do you have? What token controls do you have? What is the efficiency of your models?\u2019\u201d<\/p>\n It\u2019s against this backdrop that the Linux Foundation this week unveiled plans for the Tokenomics Foundation, a new standards body that aims to instill the same cost discipline around AI tokens that FinOps did for cloud spend. <\/p>\n \u201cIn April and May, I started hearing from companies: \u2018Oh my god, we are 3x over our entire 2026 token budget and it\u2019s only April,\u2019\u201d J.R. Storment, executive director of the FinOps Foundation, a project under the Linux Foundation, told TechCrunch. \u201cWe started hearing existential crises, and the whole conversation shifted from tokenmaxxing<\/a> and \u2018go fast\u2019 to \u2018we need guardrails, how do we control this?\u2019\u201d<\/p>\n The cries heard round the tech world followed fervent demands from CEOs pushing their teams to use the best models and move fast, costs be damned. New models released in November like Anthropic\u2019s Claude Opus 4.5, OpenAI\u2019s GPT-5.1, and Google\u2019s Gemini 3 Pro brought significant improvements to agentic tools, which have multiplied consumption. It\u2019s how one company reportedly<\/a> found itself with a $500 million Claude bill after forgetting to set usage limits for employees.\u00a0<\/p>\n \u201cIt\u2019s like the crack-cocaine epidemic,\u201d said Chris Reed, senior director of IT finance at Priceline, noting the company had begun placing token limits on certain groups. \u201cThey let you try it to get you hooked on it, and now you\u2019re kind of beholden to it.\u201d<\/p>\n Vitaly Gordon, CEO of engineering operations platform Faros AI, said he recently spoke to a CTO who told him: \u201cOne of my engineers spent $40,000 on tokens last month, and I genuinely don\u2019t know whether I should stop him or should I go and tell everyone else to be like him.\u201c <\/p>\n A March survey<\/a> by Faros found that among 20,000 developers, output was rising, but so were bugs and rewrites. Jellyfish, an engineering management platform, similarly found engineers who used the most tokens were about twice as productive than those who used AI less, but they spent 10x the number of tokens to get there. <\/p>\n Nicholas Arcolano, head of research at Jellyfish, told TechCrunch via email that expenditure on AI is exploding in large part due to agentic features, with per-developer consumption rising about 18.6x in nine months. All in all, these stats make the productivity case murkier than the spending suggests.<\/p>\n \u201cWhether extreme spend pays off comes down to the ultimate business value of shipped code (e.g. revenue), which most companies still can\u2019t measure,\u201d Arcolano said.<\/p>\n At least some of that measurement issue is the sheer scale at which AI is being used today. <\/p>\n \u201cTracking cloud costs is a hundreds-of-millions-of-rows-a-month data problem,\u201d Storment said. \u201cTracking token costs is a trillions-of-rows-a-month data problem. You can\u2019t just stick that into whatever spreadsheet or even basic tool. You\u2019ve got to fundamentally rethink your tooling, your specs and your accounting systems to do that.\u201d<\/p>\n At Priceline, Reed is already seeing discrepancies. He noted issues between a vendor\u2019s reported usage and Priceline\u2019s internal data.<\/p>\n \u201cI started my career in telecom expense management, and I\u2019m seeing all the same parallels, from telecom to cloud to AI,\u201d he said. \u201cAnytime you introduce something new, it\u2019s ripe for billing errors and audit and optimization opportunities.\u201d<\/p>\n A market is beginning to form around this problem. There are the pure-play companies, like Pay-i, which tracks, measures and optimizes the costs and performance of GenAI investments. Paid<\/a>, meanwhile, lets developers track costs, measure usage and bill users based on actual value rather than subscription fees. <\/p>\n Then there are companies like Jellyfish, Waydev and Faros AI, which all provide AI agent monitoring to prove the ROI of developer tools. Storment says most of the 180 vendors within the FinOps Foundation are leaning towards this space.\u00a0<\/p>\n Companies with existing distribution are also adding new features to capitalize on this new market. Ramp has recently moved into AI spend management<\/a>; Datadog<\/a> and New Relic<\/a> have tacked on services like cloud cost management, token-level observability, and GPU monitoring. At the FinOps X conference next week, AWS is expected to introduce new financial management features geared toward enterprise AI spending.<\/p>\n Tiffany Luck, a partner at NEA, thinks token efficiency and observability will likely be added in at the \u201charness or app layer.\u201d She pointed to Factory, a startup<\/a> that makes AI agents for enterprises, which this week launched<\/a> a model router that automatically picks the right model for every task.\u00a0<\/p>\n Gordon expects frontier labs and other model providers to adopt OpenRouter-style optimization to drive queries to the cheapest models \u2014 a trend already showing up on enterprise Claude bills.\u00a0<\/p>\n \u201cThe financial report for how much you spend on Anthropic, even if you call the Opus model, some of the spend will be on Sonnet or Haiku, because they are smart enough to do it,\u201d Gordan said. \u201cI think this will become more and more of a thing.\u201d<\/p>\n But all these tools are being built without a common language or shared definitions for how much a token costs, what it produces, and how to compare spend across vendors. That\u2019s where the Tokenomics Foundation hopes to prove useful.<\/p>\n The Foundation is building a canonical definition and framework for \u201ctokenomics;\u201d open standards, specifications and metrics for AI token usage and billing; as well as new metrics for AI economics, like cost-per-intelligence or tokens-per-watt. It also plans to define metrics across token factory effectiveness and consumption efficiency. The group is planning a formal launch in July, and is about to announce more members at the FinOps X conference next week.\u00a0<\/p>\n \u201cToken economics is fundamentally more abstract and opaque than anything we\u2019ve managed at this scale before,\u201d Nishant Gupta, chief availability officer at Salesforce, said in a statement. \u201cIt requires a different operational muscle than the one the industry built for cloud.\u201d<\/p>\n That said, Goldman Sachs projects<\/a> global token usage to multiply by 24 times by 2030. The companies already over budget need solutions now, and the foundation\u2019s first deliverable is still months away.<\/p>\n \u201cMaybe we created a steam engine, but we still haven\u2019t figured out the assembly line,\u201d said Gordon.<\/p>\n According to Arcolano, the smart move is broad, moderate adoption.\u00a0<\/p>\n \u201cThe best ROI comes from moving the broad middle from low to moderate usage, not pushing heavy users higher,\u201d he said.<\/p>\n Russell Brandom and Tim Fernholz contributed to this reporting.<\/em><\/p>\n<\/div>\n When you purchase through links in our articles, we may earn a small commission<\/a>. This doesn\u2019t affect our editorial independence.<\/em><\/p>\n