A newly discovered Windows botnet known as x47.c has recently been uncovered, offering a range of attack methods, including a particularly concerning one that targets draining victims’ paid AI credits. This revelation comes from research conducted by Qrator Research Labs and published on September 23. The botnet is being sold by a seller known as WraithTools, who is offering a variety of malicious tools and services.
One of the most alarming features of this botnet is the “AI API drain” command, which specifically targets AI API keys for services such as OpenAI and xAI. This command sends repeated billable requests directly to the AI provider, essentially draining the victim’s AI credits without their knowledge. This type of attack, known as denial of wallet (DoW), can have significant financial repercussions for the victim. What makes this attack even more insidious is that it bypasses the victim’s website, allowing it to remain operational while the AI features it relies on run out of credit.
The seller of the botnet has marketed this attack method as a way to target chatbots, AI-connected content management systems, trading bots, and scanners. They even suggest using it as a service to sabotage competitors. The seller also recommends enabling automatic top-ups to ensure that charges continue to accrue once the victim’s balance runs out. The cost of this attack varies depending on the victim’s spending limits and the extent of the extra charges.
In addition to the AI credit draining method, the x47.c botnet offers a range of other attack methods, including DDoS attacks, proxying, and AI-assisted persistence. The botnet also includes a “AI Stealth” module that uses xAI’s Grok to assess infected hosts and take predefined actions to maintain persistence and concealment. The botnet is also capable of stealing browser passwords, cookies, and Discord tokens, as well as turning infected machines into SOCKS5 proxies.
The botnet was being sold for prices ranging from $200 to $950, with the higher-priced packages including additional features such as credential theft, proxying, and AI-assisted persistence. In response to this threat, Qrator recommends revoking any exposed AI keys, monitoring billing for unusual activity, setting spending limits and controls on automatic top-ups, and implementing endpoint cleanup and DDoS protection measures.
In conclusion, the x47.c botnet poses a significant threat to Windows users, particularly those who rely on AI services. It is crucial for individuals and organizations to be vigilant and take proactive measures to protect themselves against this type of malicious activity.
