Source-linked reports: Amounts and outcomes come from the original reports and linked replies.
gdxxozn: paid AI Gateway Unified Billing credits that never produced an inference.
Community user gdxxozn says Cloudflare charged for AI Gateway Unified Billing credits, but managed Anthropic calls fail with HTTP 401 and an OpenAI call with HTTP 402, so the balance is unused. The author asks for the entitlement to be provisioned or the unused credits refunded.
What happened
- Reported cause
- Unknown
- Incident date
- Date not recorded
- First reported
- 2026-07-26
- Billing outcome
- Request submittedAuthor-reported
- Case progress
- Not established
- Public Cloudflare reply
- Not linked
Latest reported outcome
The author later reports opening Billing case #02262327, which references an earlier case and asks for escalation to AI Gateway engineering and Billing.
View this update in the timelineOriginal sources
Each entry links to the report or reply it describes.
Case timeline
Author reports the disputed charge
Open original sourceThe author says the support portal sign-in fails with ErrorCode 41 and that the earlier technical evidence is in a separate topic. The purchase amount is not stated in the post.
More to read
Why hard spending caps take time: Ashley Peacock on billing latency
Ashley Peacock (@ashleypeacock), who works on Durable Objects at Cloudflare, explains from experience why near-real-time spending caps are hard to build, says hard billing caps are coming soon, and suggests budget alerts plus issue alerts in the meantime.
RetainDB: a reported $36k bill from a queue loop with 81 users.
Reddit user alameenswe, building RetainDB on Workers, KV, Durable Objects and Queues, reports a Cloudflare bill of about $36k for one month with 81 users; the post body says $35,000. He attributes it to a queue consumer that re-queued async writes, 12 unbatched Durable Object writes per memory write and a kv.list() authentication fallback.
A budget alert is a horn, not a brake
@GUYUE_LGY points to Cloudflare’s Budget alerts documentation: alerts send email when spend crosses a threshold but do not pause or cap usage. Before letting AI run unattended, the author checks what actually happens at the line.