The ghost of a 19th century English economist
may be haunting yet another part of the AI boom.
In 1865, William Stanley Jevons observed that when the Watt steam engine made coal use more efficient—decreasing the amount required for a task—coal consumption actually skyrocketed.
More than 150 years later, one economist is citing this phenomenon, dubbed
Jevons paradox, to explain why the cost of AI will continue to creep up.
Despite the price of a single token dropping more than 90% since 2023, spending on large language models has doubled since late last year, according to the
Silicon Data Token Expenditure Index.
Apollo chief economist Torsten Slok said it’s yet another example of the Jevons paradox in action. “As tokens get cheaper, companies don’t spend less but instead run more AI agents, automate more workflows and generate more code, pushing aggregate expenditure higher even as the unit cost of intelligence collapses,” Slok
wrote in a recent blog post.
The cost of tokens has become a major concern for companies racing to leverage AI. The trend of “tokenmaxxing,” in which employees blitz to increase their AI use, has emerged as companies like Meta and Amazon incentivize the technology’s use.
But the tactic is proving unsustainable. Uber president and COO Andrew Macdonald recently said the rideshare company
burned through its entire AI budget in the first four months of the year, leading it to
cap monthly AI spending at $1,500 per employee.
In a brief last week, Bain and Co. analysts found that while token costs were halved from December 2024 to December 2025, the tokens consumed grew by 450% in the same period.
Companies feel compelled to upgrade their AI models to take advantage of the upgraded technology, they wrote, rather than stick with their current models and pocket the savings.
“The models get cheaper. The usage gets heavier,” they wrote. “The bill stays stubbornly high.”
—Sasha Rogelberg
