A corporate fad of "tokenmaxxing" on artificial intelligence technology is hitting its limits as workplaces throwing AI at everything are seeing costs rise without a similar spike in productivity, according to a report by the Associated Press.
What began as tech-industry-fueled hype in spring 2026 over squeezing as much AI-generated work as possible out of products like OpenAI's ChatGPT and Anthropic's Claude has shifted to a summertime backlash. Executives, analysts, and developers are now openly questioning whether maximizing AI token consumption was ever a sound strategy. For ongoing coverage of the AI business landscape, visit our latest AI industry reporting.
The Rise and Fall of a Fad
"Tokenmaxxing" refers to maximizing the usage of tokens — the building blocks of generative AI that correspond to small pieces of text that an AI system reads or writes. Each token represents roughly three-quarters of a word, and AI providers typically impose usage limits, with more expensive subscriptions offering higher caps.
The trend was actively promoted by some of technology's most prominent leaders. OpenAI CEO Sam Altman said in May 2026 that he was "excited to see what will happen with tokenmaxxing startups, both for how they work internally and the products they can build." Nvidia CEO Jensen Huang declared that "if your $500K engineer isn't burning $250K in tokens, something is wrong." Facebook parent Meta ran an internal competition rewarding high token usage.
But as costs accumulated, enthusiasm waned. "As bills started to pile in, people realized that those new tools are quite expensive and you need to use them wisely," said Vincent Gusdorf, head of AI analytics at Moody's Ratings and author of a new report recommending a more disciplined approach.
"It's very easy to create something you don't need with AI," Gusdorf added.
Executive Backlash
The backlash has been led by some of the tech industry's most influential voices. Microsoft CEO Satya Nadella admitted that tokenmaxxing can be addictive but warned that customers are effectively paying twice for AI — first in spending on tokens, and second by feeding their proprietary data to AI providers. Nadella's comments were unusual in the way he raised doubts about the data protection assurances of leading AI companies, even as he promoted Microsoft's own AI strategy.
Palantir CEO Alex Karp went further, telling CNBC that something had gone "completely wrong." He said he was channeling the voice of American businesses privately "livid" about paying enormous sums for tokens that create no value.
"The basic view among enterprises in this country is, 'I'm going to chillax and waste my time with tokens. I'm going to get no value and they're going to get my IP,'" Karp said.
Mozilla's chief technology officer, Raffi Krikorian, compared tokenmaxxing to the discredited practice of measuring programmer productivity by lines of code. "I think tokenmaxxing is moving through the exact same pattern," he said. "I think this is going to be an interesting blip that we're all going to look back to laugh at in a year."
The Cost Problem at Scale
The financial impact becomes stark at enterprise scale. Bain & Company management consultant Jue Wang said many of the big businesses her firm advises have been taking a closer look at returns on their AI investments.
"The token cost for them has been doubling, almost every other month," Wang said. "Let's say $200 per developer per month. Multiply that by 20,000 developers, which is often what we're dealing with at these companies, and that quickly gets you to a number that is not a line item that any general manager has planned for."
Wang pointed to a common misuse pattern: defaulting to the most powerful — and most expensive — AI models for tasks that do not require them. "Not everything needs a Claude Opus 4.6," she said of one of Anthropic's most capable models, suited to software engineering or deep research. "And yet you see so many companies, so many users, default to using Opus for everything, including generating emails."
The Rise of Model Routing and Open Weights
The cost pressure has fueled demand for AI "model routing" tools that automatically send simpler queries to cheaper, more efficient AI systems while reserving powerful models for complex tasks. Software developer Hassan El Mghari, who leads developer experience at the startup Together AI, said companies' sticker shock over the "ridiculous amount of money" spent on subscriptions to AI products from leading U.S. companies has led many to rethink the maximize-usage approach.
"It's better to kind of just empower employees on how to use this stuff and let them use AI when and however much they need to," El Mghari said.
Meanwhile, proponents of high token consumption are finding new ammunition in open-weight models from Chinese startups like Moonshot's Kimi and Zhipu's GLM, which nearly match the capabilities of top U.S. models at a fraction of the cost. Krikorian acknowledged there is "some validity to the theory that this could push tokenmaxxing a little bit further," but maintained that the broader industry trend is toward more disciplined use.
Lessons for the Enterprise
The tokenmaxxing episode offers a cautionary tale for businesses navigating the AI landscape. While leading AI model developers like OpenAI and Anthropic benefited from the surge in consumption, the experience has prompted many organizations to adopt more measured approaches — focusing on specific use cases where AI delivers measurable returns rather than maximizing raw usage metrics.
As the initial AI hype cycle gives way to pragmatic implementation, the companies that succeed may be those that treat AI as a precision tool rather than a blunt instrument, matching the right model to the right task and keeping a close eye on the bill.
