AI Costs Spiral Out of Control: Companies on the Brink of Crisis

In the tech world, major companies are forced to rethink their budgets as artificial intelligence (AI) costs rise unexpectedly. For instance, Uber exhausted its AI coding budget for 2026 by April of this year. Microsoft revoked Claude Code licenses provided to its developers after just a few months. Despite lower token prices, the proliferation of autonomous agents has sharply increased consumption volume. Techcrunch.com reports .

Although companies started 2025 relying on unlimited subscriptions, they are now trying to understand where funds are going and curb expenses. According to Alexander Embricos, head of OpenAI's corporate division, conversations with clients have shifted from model capabilities to efficiency and token control. Market participants are now developing new tools and standards to monitor AI costs.

The Tokenomics Foundation project, announced by the Linux Foundation, aims to address this very issue. J.R. Storment, Executive Director of the FinOps Foundation, noted that many companies had already spent three times their annual budget by April. This situation has triggered a genuine existential crisis in the tech sector, replacing the "move fast" mantra with "guardrails and control."

While next-generation models like Anthropic's Claude Opus 4.5, OpenAI's GPT-5.1, and Google's Gemini 3 Pro have enhanced agentic functions, this has led to a multiplicative increase in consumption. According to some reports, one company was forced to pay $500 million for Claude because it forgot to set limits for employees. Priceline representatives compare this situation to a unique form of "addiction."

Research conducted by platforms like Faros AI and Jellyfish shows that engineers who use AI tools extensively can be twice as productive, but they consume ten times more tokens to achieve this. At the same time, the number of code errors and rewrites is increasing, casting doubt on the economic efficiency of artificial intelligence.