Tech companies that embraced artificial intelligence spending in early 2025 are now facing a reckoning: token costs have exploded so dramatically that some organizations are cutting budgets, revoking access, and scrambling to understand where their money went. Uber burned through its entire 2026 AI coding budget by April, Microsoft revoked developer licenses for Anthropic's Claude, and one company reportedly discovered a $500 million bill after forgetting to set usage limits.

What's Driving the Sudden Spike in AI Costs?

The problem isn't that AI has gotten more expensive per token. In fact, per-token prices have fallen across the industry. Instead, the explosion stems from two converging forces: companies are pushing harder to adopt AI tools, and increasingly autonomous agents are consuming vastly more tokens to complete tasks.

When new models launched in November, including Anthropic's Claude Opus 4.5, OpenAI's GPT-5.1, and Google's Gemini 3 Pro, they brought significant improvements to agentic tools, which multiplied token consumption. The result has been staggering. According to research by Jellyfish, an engineering management platform, per-developer token consumption rose approximately 18.6 times in just nine months.

The productivity gains, however, don't match the spending increases. A March survey by Faros AI found that among 20,000 developers, output was rising, but so were bugs and rewrites. Jellyfish found that engineers who used the most tokens were about twice as productive as those who used AI less, but they spent 10 times the number of tokens to get there.

"One of my engineers spent $40,000 on tokens last month, and I genuinely don't know whether I should stop him or should I go and tell everyone else to be like him," said Vitaly Gordon, CEO of engineering operations platform Faros AI, recounting a conversation with a chief technology officer.

Vitaly Gordon, CEO at Faros AI

How Are Companies Trying to Regain Control?

The shift in conversation among enterprise leaders has been dramatic. Alexander Embiricos, OpenAI's head of enterprise, noted that six months ago, discussions centered on capability and quality. Now, the focus has completely changed.

"Our conversations are never about that now. Now the conversations are about, 'hey, we're spending so much. What visibility do you have? What auditability do you have? What token controls do you have? What is the efficiency of your models?'" said Embiricos.

Alexander Embiricos, Head of Enterprise at OpenAI

Companies are implementing several strategies to manage runaway costs:

• Token Limits: Priceline has begun placing token limits on certain groups of employees to prevent unlimited spending on AI tools.
• Model Routing: Some enterprises are adopting optimization strategies that automatically route queries to cheaper models, similar to OpenRouter's approach, ensuring that even when developers call expensive models like Claude Opus, some queries are fulfilled by cheaper alternatives like Sonnet or Haiku.
• Cost Visibility Tools: A market is forming around AI spend management, with startups like Pay-i and Paid offering tracking, measurement, and optimization of AI costs, while established vendors like Ramp, Datadog, and New Relic are adding token-level observability features.

Chris Reed, senior director of IT finance at Priceline, drew a striking parallel to his earlier career: "I started my career in telecom expense management, and I'm seeing all the same parallels, from telecom to cloud to AI. Anytime you introduce something new, it's ripe for billing errors and audit and optimization opportunities".

Why Is a New Industry Foundation Being Created?

The scale of the problem has prompted the Linux Foundation to unveil a new standards body called the Tokenomics Foundation, designed to bring the same cost discipline to AI tokens that FinOps (financial operations) brought to cloud spending.

The urgency is real. J.R. Storment, executive director of the FinOps Foundation, explained that in April and May 2026, companies began reporting existential budget crises. "In April and May, I started hearing from companies: 'Oh my god, we are 3x over our entire 2026 token budget and it's only April,'" Storment said. "We started hearing existential crises, and the whole conversation shifted from tokenmaxxing and 'go fast' to 'we need guardrails, how do we control this?'".

The challenge is unprecedented in scale. Tracking cloud costs involves handling hundreds of millions of rows of data per month. Tracking token costs, by contrast, involves trillions of rows per month, requiring a complete rethinking of accounting systems and tooling.

What Standards Will the Tokenomics Foundation Create?

The Tokenomics Foundation is building several critical components to address the industry's cost management crisis:

• Canonical Definitions: A shared definition and framework for "tokenomics" so companies can speak the same language about AI costs and consumption.
• Open Standards and Metrics: Specifications for AI token usage and billing, including new metrics like cost-per-intelligence and tokens-per-watt, which measure efficiency in ways cloud metrics never did.
• Consumption Efficiency Measures: Metrics across token factory effectiveness and consumption efficiency to help companies understand whether their spending is actually producing value.

Nishant Gupta, chief availability officer at Salesforce, acknowledged the challenge: "Token economics is fundamentally more abstract and opaque than anything we've managed at this scale before. It requires a different operational muscle than the one the industry built for cloud".

The Foundation is planning a formal launch in July, with more members expected to be announced at the FinOps X conference next week. However, companies already over budget need solutions now, and the foundation's first deliverable is still months away.

Goldman Sachs projects that global token usage will multiply by 24 times by 2030, meaning the cost management problem will only intensify. For now, companies are caught between the pressure to innovate with AI and the harsh reality of bills that are spiraling faster than productivity gains can justify.