The 'Tokenpocalypse' is here: Companies quietly throttle AI as token bills pile up

Promises that artificial intelligence would streamline office jobs and cut costs are colliding with a different reality as routine workplace use is running up expensive AI "tokens." 

A leaked Accenture audio clip offers a public glimpse into companies quietly trying to dial back that usage.

What happened?

Per leaked internal audio cited by 404 Media, Accenture is reportedly moving to curb employee AI use after seeing "soaring token spend." The issue is said to extend beyond engineers to nontechnical staff using AI for everyday tasks, such as turning PDFs into slide decks.

404 Media dubbed this the start of the "Tokenpocalypse."

"We're seeing from some of the data internally at least that it's actually not our engineers that are driving the token consumption. It's a lot of the non-engineers that are doing some of those behaviors [...] you were talking about," Accenture's agentic AI strategy lead Justice Kwak said in the meeting, according to the leaked audio.

The pattern may not be limited to Accenture. Some providers, including GitHub, are reportedly shifting away from flat subscriptions and toward per-token pricing, while Uber is said to have capped workers' use of AI coding tools after previously urging staff to "use AI as much as possible."

Why does it matter?

It could signal a more practical stage of the AI boom, one defined by budget discipline rather than excitement. For workers, that may translate into tighter limits, more monitoring, or narrower choices about which tools they can use. For companies, it raises a basic question: Is AI lowering costs, or just adding another operating expense?

There is also a broader infrastructure question. Every prompt, model run, and cloud query depends on power-hungry data centers, which ties AI use closely to the energy grid.

AI can still provide real advantages, including helping utilities forecast demand, integrate renewable energy, and run power systems more efficiently. But the downsides are substantial, too, including higher electricity use, heavy water demand for cooling, security and misuse risks, and the possibility that those costs could eventually reach households and businesses through higher bills.

Even companies that have pushed AI adoption are now trying to put guardrails around the cost.

Get TCD's free newsletters for easy tips, smart advice, and a chance to earn $5,000 toward home upgrades. To see more stories like this one, change your Google preferences here.