Uber president says AI spending is getting ‘harder to justify’

The specific detail worth tracking is that Uber burned through its annual AI budget before mid-year, which is less a story about frugality and more a signal that enterprise token consumption is scaling faster than procurement teams modeled, creating a structural mismatch between contract cycles and actual usage curves.

This sits directly alongside the Stratechery piece on Nvidia's bifurcated reporting from the same week, which flagged margin compression and commoditization pressure at the hyperscaler tier. Uber's budget problem is the demand-side mirror of that supply-side story: infrastructure spending is accelerating, but the enterprise customers funding it are hitting internal justification walls before vendors hit capacity limits. The WIRED piece on AI agents and Claude Code's role in destabilizing developer workflows adds another layer, since Claude Code is the specific product Uber is struggling to cost-justify. Together, these three stories sketch a coherent picture where agent-driven token consumption is outpacing both enterprise budgets and demonstrable productivity gains.

Watch whether Anthropic responds with enterprise pricing restructuring or usage-cap tiers for Claude Code within the next two quarters. If they do, it confirms that Uber's situation is widespread enough to threaten retention, not an isolated edge case.