Uber burned its 2026 AI coding budget by April; 35% of big enterprises are cutting token spend
ruslansv · x · 2026-11-16
Ruslan Belkin examines enterprise AI coding budgets. Per The Information, Uber exhausted its full 2026 AI coding budget by April. A Sonar survey of 1,026 developers (released Nov 12) found 35% of respondents at 1,000+ employee companies had cut back AI coding to control token spending, vs 21% at companies with 200 or fewer employees.
His argument: a generous initial budget is right—engineers need room to experiment—but the problem starts when experiments become recurring production expenses without a decision about what they should buy. Across-the-board cuts leave all three mistakes intact: model choice, caching, and routing can shrink the bill, and a concentrated rewrite can make a larger bill worthwhile. The next budget should reflect what was learned, not simply how much was spent.
On the other side, Anthropic's IPO filings show nearly $17B in Q3 2026 revenue, over 12x the year-earlier quarter.
More from coding & agent
- Vibe Check: Gemini 4 Carbon Codes Better Than Argon but Trails Fable 5.2 on Judgment — every · 2026-11-17
- Gemini Managed Agents Adds Self-Hosted Sandboxes: Google Runs the Loop, You Run the Tools — _philschmid · 2026-11-17
- Gemini Managed Agents now run bash and file edits in your own compute — _philschmid · 2026-11-17
- How to Pick Between Gemini 4 Carbon and Hydrogen: Plan With Carbon, Execute With Hydrogen — _philschmid · 2026-11-17
- Dev builds Simple Listen: turns articles and AI agent output into a private podcast feed — haltakov · 2026-11-17
- Agent vibe-codes a custom bootable app for Ableton Push 3 — kieranklaassen · 2026-11-17