Microsoft Cancels Internal Anthropic Licenses As Shift To Token-Based AI Billing Blows Up Annual Budgets In Months

The good news for AI generators and investors is that the technology is beginning to prove out financially as revenues surge. The bad news? The industry's move from flat-rate fees to token-based billing that charges for every line of code generated has become so expensive that it is blowing up annual budgets in a matter of months. Uber has warned employees that it has burned through its entire 2026 budget in four months. Microsoft, which has infinite cloud-computing assets, has cancelled a program providing internal staff access to Anthropic's Claude because of the unexpectedly exorbitant cost. 

The implication is that the tech industry is trying to cash in in order to optimize its profits - and help pay for some of the investments it has made - to such an extent that even its ecosystem partners are pulling out. It is prudent to assume that a modus vivendi will eventually be reached that allows AI companies to recoup some of their investments without killing usage. But when that will be and how it will be engineered is not yet apparent. In the meantime, the push and pull between provider and customer could suppress some demand. JL

Blaze Trends Tech reports:

The enterprise AI gold rush hit a very expensive wall. Companies are realizing that giving thousands of employees unlimited access to advanced AI models is blowing through yearly budgets in a matter of months. Providers are moving away from predictable flat-rate fees, shifting to expensive token-based usage models that charge for every prompt and line of code generated. The shift toward token-based billing is causing intense budget overruns. Uber sent out an internal memo warning that the company burned through its entire 2026 AI budget in just four months. US AI software prices have jumped between 20% and 37%. That cost reality just forced Microsoft to begin canceling internal licenses for Claude Code, ending a pilot program that gave its staff access to the AI tool. When even Microsoft—a company with infinite cloud computing resources—decides a third-party coding assistant is too expensive for its own staff on a usage basis, it sets a precedent.

The enterprise AI gold rush is suddenly hitting a very expensive wall. Across the software industry, companies are realizing that giving thousands of employees unlimited access to advanced AI models is blowing through yearly budgets in a matter of months. Providers are moving away from predictable flat-rate fees, shifting to expensive token-based usage models that charge for every prompt and line of code generated.

That cost reality just forced a major internal pivot at one of Anthropic’s biggest partners. Microsoft has begun canceling internal licenses for Claude Code, abruptly ending a pilot program that gave its developers and non-technical staff access to the AI-powered programming tool.

The initiative originally kicked off in December 2025. Thousands of Microsoft employees were invited to experiment with the AI assistant to speed up their workflows. But the enthusiasm of that initial testing phase quickly collided with the friction of enterprise production viability. 

The company is officially concluding the Claude Code pilot program, pulling access from staff who had integrated the tool into their daily routines. Basic AI access remains relatively affordable for consumers. But the price of running advanced, enterprise-grade models at scale has surged. Data from Tropic shows that American AI software prices have jumped between 20% and 37% recently.

The Token Economy Burns Through Budgets

Microsoft is far from the only giant feeling the burn of these rising API costs. The shift toward token-based billing is causing intense budget overruns across Silicon Valley. Uber’s CTO, Praveen Neppalli Naga, sent out an internal memo warning that the ride-hailing company burned through its entire 2026 AI budget in just four months. 

When you pay by the token, a few heavy users running complex code generation tasks can drain millions of dollars faster than finance departments can track them. 

Anthropic itself recently modified its enterprise billing framework for Claude models to reflect this new reality. And the pressure is coming from inside the house at Microsoft, too. The company’s own code repository platform, GitHub, announced it is abandoning flat-rate plans in favor of usage-based billing.

What the Death of Flat-Rate AI Means for Enterprise Software

The cancellation of the Claude Code pilot at Microsoft is a direct signal that the experimental phase of enterprise AI is over. We are entering the optimization phase. For the last two years, companies threw unlimited money at AI tools just to see what would stick. Now, CFOs are looking at the token usage receipts.

When even Microsoft—a company with effectively infinite cloud computing resources—decides that a third-party coding assistant is too expensive to license for its own staff on a usage basis, it sets a precedent for the rest of the industry. 

Software vendors that rely on wrapping OpenAI or Anthropic models will have to completely rethink their pricing tiers. If they stick to flat monthly fees, power users will bankrupt them. If they pass the usage costs directly to the customer, enterprise adoption will slow down significantly.