Reference · 2026

Runaway AI Agent Spend: Failure Modes

How an autonomous agent actually runs away with money — the failure modes, the arithmetic that makes them expensive, and the caps that bound them.

By the sipi.bot engineering team · Published 2026-07-18 · Last updated 2026-07-25 · Methodology

Correction (2026-07-25). An earlier version of this page reported an incidence rate and a median incident cost, attributed to a survey of production agent teams and to aggregated customer records. No such survey was ever run, and there are no customer records to aggregate: the firewall is MIT-licensed, self-hostable and collects nothing by design. Those figures were fabricated and have been removed rather than restated. What remains below is the failure-mode taxonomy and cost arithmetic, which you can verify yourself from published provider pricing.

What a runaway incident is

A runaway incident is an agent executing an action sequence — typically a retry loop or a tool-call chain — that spends beyond what the operator intended for that session. The agent did something the operator did not intend, and it cost money.

The four failure modes

These are the recurring shapes of runaway spend. They are engineering failure modes, not survey results, and each maps to a control that stops it before the spend happens:

Failure modeWhat happensControl that bounds it
Retry loop on a failed tool callOne failing call is retried without a ceiling; each retry is billedVelocity cap (max transactions per hour)
Prompt injection to an unintended vendorInjected instructions redirect purchasing to an attacker-chosen destinationAllowlist + per-transaction cap
Unattended overnight sessionNo human is present to notice or intervene for hoursDaily total cap
Aggressive provisioning while debuggingLoosened limits during development are never tightened againEnvironment-scoped rules

The retry loop is the one worth designing for first: it needs no attacker and no unusual conditions, only a failing dependency and no ceiling.

Why the tail dominates

Runaway cost is not driven by the price of a typical task — it is driven by repetition. The arithmetic is simple enough to check against any provider's published rates:

Work the numbers for your own stack with the spend calculator and the published per-model rates in ai-model-costs-2026.csv.

Methodology

What this page is based on: published provider pricing (rates each vendor lists publicly, captured in ai-model-costs-2026.csv), publicly reported incident write-ups, and the failure modes exercised by the open-source evaluation suite — 53 scenarios you can run yourself with pip install sipi-bot.

What it is not based on: any survey, and any customer or deployment telemetry. sipi.bot collects none — no telemetry is a design property of the product, not an oversight — so it is not in a position to publish incidence rates, median incident costs, or percentile distributions, and it does not.

If you have run into runaway agent spend and can share what happened, the repo is the place — real incident write-ups are worth more than an invented percentage.

Install the firewall that prevents these incidents →