In October 2026, a large-scale attack targeted providers using VMmanager, allowing unauthorized access to ISO image mounting functions. Attackers executed commands on panel servers without credentials, leading to the theft of configuration files, access keys, and crypto wallets. Companies like FirstVDS, Senko Digital, and MHost were affected by the incident.
The scale and synchronization of these attacks suggest the use of automated tools. Experts note that AI significantly accelerates the development and adaptation of exploits. While direct evidence of AI usage in the VMmanager breach is absent, the industry is seeing a trend where LLMs are used to scan infrastructure software for vulnerabilities.
Recent global incidents confirm the efficiency of AI-driven attacks. Autonomous agents successfully targeted over 1,200 websites across the US, Japan, Australia, and Canada, with successful breaches costing approximately $50 and taking about 23 hours. Another case involved a novice hacker compromising 14 companies using Claude and Codex.
Modern cybersecurity faces a new reality where attack 'breakout times' have dropped significantly. According to CrowdStrike and ReliaQuest, criminal groups can now move between machines within a network in a matter of minutes, with record times as low as 27 seconds. Traditional defense methods are struggling to keep pace with this automated threat landscape.
