Chinese Hacker Breaches Deere, Cummins Using $5,000 AI Agent Fleet

A Chinese-speaking cybercriminal breached websites belonging to more than 100 companies, including agricultural giant John Deere, appliance maker SharkNinja, and engine manufacturer Cummins, starting September 25. Snorre Fagerland, principal threat researcher at cybersecurity firm NorForge, identified the intrusions by scanning the web for exposed instances of Cairn, an open-source AI problem-solving platform. The hacker did not manually exploit vulnerabilities; instead, they deployed AI agents that autonomously probed, identified weaknesses, and executed intrusions over a period of weeks. The campaign targeted e-commerce sites in the US, Canada, Australia, and Japan, with the hacker’s instructions limited to fewer than 100 Chinese characters describing preferred tactics.

The efficiency of the attack was stark. Fagerland estimated the entire operation consumed roughly $5,000 worth of AI tokens, averaging about $50 per successful hack. The agents typically required 23 hours of continuous, unattended work to break into a target. In one notable instance, after failing to find vulnerabilities in a primary application, the AI pivoted to search digital records for associated infrastructure, located a public-facing control panel, and exploited a subtle data processing flaw to execute malicious commands. Fagerland described the technique as sophisticated, noting, “It’s not script kiddie stuff.”

Victims responded with varying degrees of transparency. Deere told Reuters it had taken “necessary remediation actions” and saw “no indication that sensitive information was accessed.” SharkNinja stated the breach hit a standalone support site and was contained. Cummins confirmed it was “investigating the report of a potential cyber-attack on a website operated by one of our third-party vendors.” Two smaller businesses also acknowledged compromise, while other victims listed in the Cairn activity logs did not respond to inquiries. Reuters could not determine if data was exfiltrated in these breaches.

The incident highlights a growing anxiety over autonomous AI capabilities in the cybercriminal underworld. Fagerland, who has worked in cybersecurity for nearly three decades, called it a “new world” where commercially available agents can pull off complex intrusions from start to finish without human intervention. The author of the Cairn platform, using the handle “le1xia0,” disputed the intent behind the tool’s use, stating in a recent X post that the architecture “was not designed for hacking attacks.” Meanwhile, South Korea and Japan have recently faced similar disruptions attributed to AI-using cybercriminals, suggesting a broader trend in the weaponization of agentic AI.