AI token consumption is exploding faster than companies can manage their budgets. OpenClaw creator Peter Steinberger recently revealed that his team spent $1.3 million on OpenAI API tokens in a single month while consuming 603 billion tokens, underscoring a dramatic shift in how organizations are grappling with the true cost of deploying advanced AI systems at scale.

This disclosure comes as OpenAI CEO Sam Altman has begun openly acknowledging a problem that barely existed six months ago: companies are exhausting their entire annual AI budgets within the first quarter of the year. The rapid acceleration of AI spending reflects both the growing power of AI systems and the unexpected scale at which organizations are deploying them, particularly through agent-based systems that can perform multiple tasks autonomously.

Why Are AI Costs Spiraling So Quickly?

The root cause is a phenomenon economists call Jevons paradox, where lower costs and greater efficiency lead to increased consumption rather than reduced demand. As AI models become cheaper and more capable, organizations are deploying them more widely across their operations. Altman painted a striking picture of this trend during a recent talk at the Intelligence at Work event, noting that OpenAI's largest customer from six and a half years ago used around 100,000 tokens per month. Today, that figure is roughly equivalent to average global per-capita token consumption, while the company's highest-volume user consumes about 100 billion tokens each month.

The scale of consumption has caught even industry leaders off guard. Altman acknowledged that another user consumes even more than that, saying he was surprised by the magnitude of the usage. This explosive growth in token consumption reflects what economists describe as Jevons paradox, where lower costs and greater efficiency lead to increased consumption rather than reduced demand.

What Are Companies Telling OpenAI About Their Budget Crisis?

The conversation between AI vendors and their enterprise customers has shifted dramatically. Altman explained the new reality: "People are really saying, you know, it's kind of a meme now, but 'My company spent my entire 2026 budget in Q1. Can you make this more efficient?' We are continuing to push on that more with models. I think we'll have a lot of ways we can help people get more value for less spend. But that went from, at the beginning of this year, an issue that never came up (people were totally happy with the amount they were spending) to, all of a sudden, a huge issue".

Altman

"People are really saying, you know, it's kind of a meme now, but 'My company spent my entire 2026 budget in Q1. Can you make this more efficient?'" said Sam Altman.

Sam Altman, CEO at OpenAI

This represents a fundamental shift in how enterprises view AI spending. Earlier in 2026, companies were primarily focused on expanding their use of AI tools and agents to improve productivity and automate tasks. However, rising usage has led to larger-than-expected AI bills for many organizations, forcing them to reassess their strategies.

How to Manage Growing AI Token Costs

• Monitor Token Consumption Closely: Track how many tokens your organization is using across all AI applications and set up alerts when consumption approaches budget thresholds. This visibility is essential for preventing surprise overages.
• Prioritize High-Impact Use Cases: Focus AI deployment on tasks where the return on investment is clear and measurable, rather than deploying agents for low-priority work that may inflate costs without corresponding benefits.
• Evaluate Cost-Benefit Trade-offs: Assess whether operating advanced AI systems is actually more cost-effective than employing human workers for comparable tasks, as some companies have found that AI can be more expensive in certain scenarios.
• Demand Efficiency Improvements from Vendors: Push AI providers to optimize their models for lower token consumption without sacrificing quality, as OpenAI and other vendors are now prioritizing efficiency gains.

The challenge facing organizations is not theoretical. Some companies have found that operating advanced AI systems can, in certain cases, cost more than employing human workers for comparable tasks. Meanwhile, reports suggest some organizations are beginning to scrutinize those costs more closely. Amazon employees have reportedly used AI agents for low-priority tasks to improve their rankings on internal AI leaderboards, while Microsoft is said to have reduced some Claude Code licenses due to rising expenses.

Even prominent executives are questioning whether aggressive AI spending translates to real business value. Uber CEO Dara Khosrowshahi has stated that there is not yet a clear link between aggressive AI spending and successful product outcomes, suggesting that the current spending surge may not be sustainable or justified.

OpenAI is now focusing on improving model efficiency as customers seek ways to balance expanding AI adoption with growing operational costs. Altman believes AI usage will continue expanding, but the company recognizes that efficiency gains must keep pace with growing demand. The challenge ahead will be whether AI vendors can deliver the cost reductions that enterprises are now demanding, or whether the current wave of AI adoption will slow as budgets tighten and return on investment becomes harder to justify.