AI Tokens Run Amok: How Companies Are Battling Skyrocketing Costs

Across the tech industry, companies are reeling from the unexpected financial hit of artificial intelligence (AI) token usage. Early adopters who gorged on all-you-can-eat subscriptions are now facing steep bills, prompting a scramble for better cost management tools.

Uber, for instance, blew through its entire 2026 AI coding budget by April, while Microsoft had to revoke developer access due to excessive token consumption. A Priceline employee told TechCrunch that routine contract renewals are now 4-5 times more expensive than expected.

As per-token prices have fallen, the push for more AI adoption and increasingly autonomous agents has driven overall token usage higher. 'Six months ago, our conversations were all about whether it was good enough,' said Alexander Embiricos, head of enterprise at OpenAI. 'Now they’re about visibility, auditability, and efficiency.'

This backdrop led the Linux Foundation to unveil plans for a new Tokenomics Foundation, aiming to standardize cost discipline around AI tokens as FinOps did for cloud spend.

Companies Are Over Budget Already

In April and May alone, many companies found themselves 3x over their entire 2026 token budget. CTOs are now asking tough questions: 'Should I stop my engineers or encourage them to use more AI?'

Key Stats and Trends

A two-year study of 20,000 developers by Faros AI found that those using the most tokens were about twice as productive but spent 10x the number. Nicholas Arcolano, head of research at Jellyfish, noted a rise in per-developer consumption of about 18.6x over nine months.

Companies are now rethinking their tooling and accounting systems to track token costs more effectively. 'Tracking cloud costs is already a huge data problem,' said J.R. Storment, executive director of the FinOps Foundation. 'Token costs are a trillions-of-rows-a-month data problem.'

New Market for Cost Management

A market has emerged to help companies manage these costs better. Companies like Pay-i track and optimize GenAI investments, while Paid lets developers track costs based on actual value.

Storment also highlighted the shift in focus from just using AI to ensuring it's cost-effective: 'We're moving away from tokenmaxxing and go fast to guardrails.'

The Tokenomics Foundation

To address these challenges, the Tokenomics Foundation is working on a canonical definition for 'tokenomics,' open standards for AI token usage and billing, and new metrics like cost-per-intelligence or tokens-per-watt.

With Goldman Sachs projecting global token usage to multiply by 24 times by 2030, companies need solutions now. The foundation's first deliverable is months away, but it's not too late for moderate adoption: 'The best ROI comes from moving the broad middle,' said Faros AI's Nicholas Arcolano.