Description

A profit-driven threat actor leveraged several commercial generative AI platforms to breach more than 600 FortiGate devices across 55+ countries between January 11 and February 18, 2026. The operation demonstrates how AI-enabled tooling can significantly lower the skill threshold required to conduct large-scale cyber intrusions, enabling smaller or less-experienced actors to operate at volume and speed. The attacker gained entry by targeting internet-exposed FortiGate management interfaces using credential-based techniques. By scanning commonly used management ports—443, 8443, 10443, and 4443—the actor identified systems secured with weak, reused, or single-factor authentication credentials. No zero-day exploits were necessary. After successful access, configuration files were downloaded, exposing sensitive data such as SSL-VPN usernames and passwords, administrative credentials, IPsec VPN configurations, firewall rules, and detailed internal network architecture. The extracted configurations were processed and decrypted using AI-supported Python scripts, allowing the actor to efficiently organize and reuse harvested credentials at scale. The campaign appeared opportunistic, relying on automated mass scanning rather than industry-specific targeting. However, in several cases, multiple devices belonging to the same organization—including managed service providers—were compromised. Affected regions included parts of South Asia, Latin America, the Caribbean, West Africa, Northern Europe, and Southeast Asia. Post-compromise activity involved deploying Meterpreter with Mimikatz to conduct DCSync operations against Active Directory, enabling the theft of NTLM credential databases. The attacker used pass-the-hash, pass-the-ticket, and NTLM relay methods for lateral movement. Backup systems, particularly Veeam Backup & Replication servers, were also targeted to disrupt potential recovery efforts. Although widespread, the campaign revealed operational weaknesses, as the actor frequently failed against well-secured environments and shifted focus rather than escalating efforts. Overall, the incident illustrates how AI is amplifying the efficiency of cybercriminal operations.