Key Takeaways
- Companies that encouraged employees to maximise AI token usage (“tokenmaxxing”) are now reversing course due to soaring costs.
- Meta, Uber, Walmart and Amazon have imposed monthly limits on AI tools and removed internal leaderboards that tracked token consumption.
- The emerging practice of “tokenminning” emphasizes using the most powerful models only for complex tasks and substituting cheaper alternatives elsewhere.
- Leaders argue that measuring AI value by output (e.g., “agentic work units”) is more meaningful than counting tokens.
- Despite cost‑cutting moves, firms still plan to spend billions on AI, but they are seeking ways to get similar or better results with lower expenditure.
The Rise of Tokenmaxxing
Earlier this year, the message from tech companies to employees was clear: Use as much artificial intelligence in your work as possible. Employees coined the term “tokenmaxxing,” with a token referring to a unit of AI use roughly equal to a word fragment. At Meta and Amazon, workers even competed on leaderboards that tracked token use, turning AI consumption into a game of volume.
Cost Shock Triggers a Reversal
Then came the bills from AI providers such as Anthropic and OpenAI—and they were not cheap. Meta told employees last week that it would soon limit AI use after seeing an “exponential increase” in costs. In May, Uber said it had blown through its projected AI spending for the year in just four months and has placed monthly limits on AI coding tools. Walmart also set limits for different AI tools, and Amazon and Meta have taken down their tokenmaxxing leaderboards.
Enter Tokenminning
In other words, “tokenminning,” short for “token minimizing,” is now in. The reversal, within just a few months, underlines how AI use remains in flux as people try to figure out how to best use the tools. “The biggest problem is this is all changing so fast, people and companies don’t know what to do,” said Rob May, the chief executive of Neurometric, a start‑up that helps companies better use AI, and the author of “The Tokenminning Manifesto.” He added that CEOs who did not know how to measure the AI savviness of their employees thought, “Well, who’s using the most tokens?”—a philosophy that ended up promoting volume over efficiency.
How AI Pricing Works
OpenAI and Anthropic offer subscriptions that cost $10 to $200 a month for use of their AI models; when subscribers hit their usage limit, they are cut off. But the bulk of the revenue comes from offering tools to companies like Meta, Shopify and Amazon, which pay not only subscription fees but also for the tokens used by their tens of thousands of workers. So the more tokens that are used, the more money the AI costs. A simple task, like asking AI to summarize the transcript from a company meeting, may use a few hundred tokens. More complex requests, like writing code to build a new product or feature, can use tens of thousands.
Escalating Costs with More Powerful Models
The costs of using AI models have soared as they have become more powerful and consume more tokens. Anthropic’s newest AI model, Fable, is twice as expensive as its previous model, Opus. While there are cheaper models, many employees have fallen into the habit of using the most powerful models for everything, Mr. May said.
Changing Patterns of AI Use
The ways that people use AI have also changed. Instead of just conversing with AI chatbots, engineers deploy AI “agents,” which can work on complex tasks for hours at a time. As a result, engineers can use tens of thousands of dollars’ worth of tokens each month. Many companies said they were trying to be more strategic about AI spending after not seeing clear returns on their investment. “If you’re not actually able to draw a direct line to how many useful features and functionality you’re shipping, that trade becomes harder to justify,” Andrew Macdonald, Uber’s chief operating officer, said in a recent podcast interview. “That link is not there yet.”
Continued Investment, New Metrics
That’s not to say companies won’t keep spending big on AI. Meta told employees that it was on track to spend billions on AI use this year, but wanted to “find places we can spend less while getting similar or better business results.” Marc Benioff, the chief executive of Salesforce, the enterprise software company, said his company planned to spend hundreds of millions on AI this year but now tracked “agentic work units” instead of tokens. The new metric is supposed to measure output, not just use.
Internal Tools and Provider Reaction
Meta’s and Walmart’s limits on employee AI use were reported earlier by The Information and Bloomberg. It’s unclear how “tokenminning” might affect the bottom lines of Anthropic and OpenAI. At the height of the tokenmaxxing era this year, the AI companies reported record revenues driven by the use of coding tools. Last week, Meta told its engineers to use its internal coding assistant, MetaCode, instead of third‑party tools if possible. Meta declined to comment, Anthropic did not provide a comment, and OpenAI did not respond to a request for comment. (The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to AI systems. They have denied the suit’s claims.)
The Path Forward: Selective Use of Cutting‑Edge AI
The clear path forward for companies, Mr. May said, is to use cutting‑edge AI only on complex tasks that require it and substitute cheaper models for other instances. Companies can save as much as 90 percent by opting for less advanced AI models, said Andy Markus, AT&T’s chief AI officer. He said his engineers were using the most powerful AI models for some tasks and the less powerful ones for most other actions. “There’s an ebb and flow,” he said. “What we do find is that, for most use cases, the latest greatest frontier model isn’t needed.”
Outlook
As the AI hype settles into a more measured phase, organisations are learning to balance ambition with accountability. The shift from tokenmaxxing to tokenminning reflects a broader maturation: AI is no longer a novelty to be chased indiscriminately, but a tool whose value must be quantified in tangible business outcomes. While the era of unrestricted token consumption may be over, the underlying drive to harness AI’s potential remains strong—only now, firms are seeking smarter, cheaper ways to get there.
https://www.nytimes.com/2026/06/18/technology/ai-token-minimizing.html