Uber burns through its entire 2026 AI coding budget in four months
Uber · Jun 2, 2026 · Coding agent
What happened
Uber exhausted its full-year 2026 budget for AI coding tools by April, four months into the year, and responded by capping employee AI spending. TechCrunch reported the overrun as part of a broader industry reckoning over runaway token costs, noting that Microsoft had similarly revoked Claude Code licenses for developers months after enabling them.
Causal vector
Uncapped per-developer token consumption across an engineering org — spend scaled with agent adoption faster than budget controls could absorb
Source
Reported by TechCrunch. Verified against the primary report.
Org-level budgets fail exactly where per-agent controls succeed: at the individual call. Daily totals and velocity limits bound every developer's agent spend at the transaction level, so a full-year budget cannot be exhausted in a quarter without an explicit decision to raise it.
The six rule types that contain this class of failure
Per-transaction cap
Any single spend above your ceiling is BLOCKED before it moves.
Daily total
Cumulative spend across all agent calls, bounded per day.
Velocity limit
Stops runaway retry loops — the #1 cause of overnight losses.
Merchant allowlist
Only approved destinations can ever receive funds.
Category rules
Flag high-risk classes (crypto, infra, refunds) for review.
Approval threshold
Above a value, the action waits for a human.
Related incidents
Company hit with a reported $500 million Claude bill after leaving usage limits off
Unnamed enterprise (reported via Axios) · May 28, 2026
OpenAI o3 ARC-AGI benchmark run estimated to cost ~$346,000+ in compute
OpenAI · Dec 20, 2024
Don't be the next entry
Every incident in this database is the result of trusting a prompt, a provider cap, or a human review cycle. sipi.bot replaces all three with one deterministic call. 67 documented failures, one control.