---
title: When AI Shuts Down a Wind Farm for $8
author: Benedikt Brünner
date: 2026-11-16
url: https://postgenerative.at/indications/2026-11-16/when-ai-shuts-down-a-wind-farm-for-8
lang: en
type: Indication
summary: In a security experiment, AI agents managed to shut down a simulated wind farm in less than two hours. The API costs for one of the attacks were just $8. The experiment illustrates how AI can lower the cost and barriers to cyberattacks on critical infrastructure.
tags:
  - Security
  - Software
sources:
  - title: AI models disable wind farm in under two hours for research purposes | DER STANDARD
    url: https://www.derstandard.at/story/3000000349319/zu-forschungszwecken-ki-modelle-legen-windpark-in-unter-zwei-stunden-lahm
  - title: AI agents bring down wind farm within two hours | Computer Weekly
    url: https://www.computerweekly.com/news/366654794/AI-agents-bring-down-wind-farm-within-two-hours
license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
alternate:
  de: https://postgenerativ.at/indizien/2026-11-16/wenn-ki-fuer-8-einen-windpark-abschaltet
---

# When AI Shuts Down a Wind Farm for $8

In a security experiment, Shinobi Security investigated how independently AI agents could attack a wind farm. The researchers used a simulated facility rather than operating infrastructure. [1][2]

Two AI models achieved the objective of disconnecting the simulated wind farm from the power grid. Claude took 66 minutes, with API costs of approximately $20. Kimi took just under two hours, with **API costs of approximately $8**. [1][2]

The agents exploited existing credentials, misconfigurations and default passwords, among other weaknesses. The experiment therefore does not necessarily reveal new vulnerabilities. Instead, it demonstrates how AI can automate and combine known attack techniques. [1][2]

**What stands out is not only the speed, but also the price:** When complex attack sequences can be automated at low API costs, an important barrier to cyberattacks on critical infrastructure is reduced.

The experiment took place under controlled conditions. It does not prove that every real wind farm could be shut down. It does, however, show how AI can exploit existing weaknesses automatically and at low cost.

**Another indication of a post-generative present in which AI systems are becoming increasingly autonomous, and our critical infrastructure must be prepared for this new reality.**
