Microsoft Banned Claude Code After 100K Engineers Adopted It; Uber Burned Through $3.4B AI Budget in Four Months
Two of the world's largest technology companies are hitting the AI cost wall simultaneously, turning token spending into the new capital expenditure crisis that CFOs never planned for.
Microsoft quietly banned Claude Code internally after roughly 100,000 engineers adopted the tool, and Uber burned through its $3.4 billion AI budget in just four months, according to a roundup by @TraffAlex. The two incidents, surfacing on the same day, paint a stark picture of enterprise AI economics in mid-2026: adoption is no longer the problem. Cost control is.
The Microsoft situation is particularly revealing. Claude Code, Anthropic's terminal-based coding agent, has become a favorite among developers for its ability to navigate large codebases, write tests, and refactor code with minimal hand-holding. That 100,000 engineers reached for it organically — without top-down procurement — suggests the tool was genuinely solving problems that Microsoft's own Copilot stack either couldn't or didn't. But at enterprise scale, every Claude Code session means tokens flowing to Anthropic's API, and 100,000 power users can generate staggering inference bills. Microsoft's decision to ban the tool rather than negotiate volume pricing or build guardrails around usage signals that the cost trajectory was unsustainable, or that the competitive optics of funding a rival's inference engine at that scale became untenable, or both.
Get our free daily newsletter
Get this article free — plus the lead story every day — delivered to your inbox.
Want every article and the full archive? Upgrade anytime.
No spam. Unsubscribe anytime.