Claude Code ran an uncapped test loop for 17 hours, racked up a $2.1K Bedrock bill while 'idle'
nymma19 · reddit · 2026-10-14
A developer reports that Claude Code proposed and ran verification tests against Claude Haiku 4.5 on AWS Bedrock as background processes. When the main session hit its usage limit that evening, the processes kept making paid calls for about 17 hours — roughly 60,000 calls and $2.1K, versus a normal daily spend of a few dollars.
Key details:
- In the same minute as the limit notice, Claude described the process as looking "idle... waiting, not looping."
- Anthropic's support bot (Fin) acknowledged "a gap between intended behavior and what occurred" and called it "a product defect," but couldn't act on Bedrock usage or escalate to a human.
- AWS says Anthropic is the seller of record, so only Anthropic can decide on a refund; the author can't reach a person at either company.
- The author has since added budget alerts and lockouts, and is asking whether others have hit this — a heavy hit for a small bootstrapped startup.
More from coding & agent
- Isolating Antigravity chats per subfolder with AGY Rules and PreInvocation Hooks — bibryam · 2026-10-16
- Gemini shown off modding Minecraft to turn players into any Spider-Man — zacharynado · 2026-10-16
- Workflow: annotate UI screenshots in tldraw, hand to Claude, get better UI — max__drake · 2026-10-16
- Beyond chat wrappers: designing stateful, multimodal AI apps as interaction conductors — dr_cintas · 2026-10-16
- IBM engineer maps 5 levels of AI coding, from manual to a 3-day agent software factory — TejasKumar_ · 2026-10-16
- 10 open-source repos that let your AI agent reverse-engineer apps and games — victor_explore · 2026-10-16