The AI industry is hitting a hard reality: the era of cheap, subsidized artificial intelligence is ending, and companies are scrambling to adapt. After months of explosive growth in AI token consumption driven by advanced reasoning models and autonomous agents, corporate customers are now facing sticker shock. What started as a goldilocks narrative of unlimited AI abundance has collided with the economics of actually running these systems at scale.

Why Are AI Costs Suddenly Becoming a Crisis?

The problem emerged quickly and publicly. Uber reportedly burned through its entire AI budget in just four months, followed by reports of a $500 million overspend at another major company. Within weeks, the conversation shifted from celebrating AI's explosive growth to what industry insiders now call "token panic".

The root cause is straightforward: advanced AI models consume vastly more tokens, the basic units of text that AI systems process, than earlier versions. When companies distributed these tools widely and encouraged their use, employees began running enormous bills without realizing the cost implications. Meanwhile, AI providers simultaneously shifted their pricing models from flat subscriptions to usage-based billing, making costs directly proportional to consumption.

According to reporting from The Economist, Anthropic's annual recurring revenue reached $45 billion in May 2026, a fivefold increase since the start of the year. This explosive growth for the AI lab translates directly into exploding operational expenses for its customers.

How Are Major AI Providers Changing Their Pricing?

The shift toward usage-based billing has been coordinated across the industry's largest players. Here's what happened in recent months:

• OpenAI (April 2): Changed Codex pricing to align with API token usage instead of per-message pricing, making costs directly tied to consumption volume.
• Google (May 19): Shifted Gemini subscriptions from "daily prompt limits" to a "compute-used" model, charging based on actual computational resources consumed.
• Microsoft (June 1): Transitioned GitHub Copilot to usage-based billing, moving away from flat subscription rates.

Adding to the confusion, some models use new tokenizers that may consume up to 35% more tokens for the same text, even though the list price remains unchanged. This hidden cost increase caught many enterprises off guard.

Sam Altman, CEO of OpenAI, acknowledged the shift in a recent statement. "Probably the second biggest theme is just around cost," Altman explained. "People are really saying, it's kind of become 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. I know. People were totally happy with the amount they were spending, to all of a sudden, a huge issue."

Sam Altman, CEO of OpenAI

Microsoft's AI Chief added blunt criticism of the pricing situation. "Anthropic is extremely expensive, and I think many people are urgently looking for alternatives," the executive stated after Microsoft cancelled Claude Code licenses in May.

What Happens When Companies Face the Real Cost of AI?

The monetization crunch is forcing a reckoning across enterprises. Companies are beginning to restrict AI functionality, invest in oversight tools to track spending, and pit AI budgets directly against hiring decisions. The question of return on investment, which has been largely theoretical during the subsidy phase, is now being answered in real time across millions of use cases.

For most users, the impact may be modest. But science projects, experimental agents, and casual AI exploration will either be cut or shifted to open-source models. The economics are shifting in favor of "good enough" solutions. As the cost of running open-source, discount, or smaller AI models continues to decline while their capabilities improve, enterprises will increasingly choose efficiency over cutting-edge performance.

This represents a fundamental shift in the AI industry's business model. For the past three years, the dominant strategy has followed the classic venture capital playbook: subsidize demand to gain market share and lock-in, then monetize once customers are dependent. The AI industry is now entering the monetization phase, whether by choice or necessity. As one analyst noted, "Eventually, customers have to start picking up the tab".

The trillion-dollar capex investment in AI infrastructure was always predicated on the assumption that revenue would eventually follow. That moment has arrived, and it's forcing both AI providers and their customers to confront the true economics of artificial intelligence at scale.