New x47.c Botnet Targets AI API Credits via Denial of Wallet Attacks

A newly discovered Windows botnet, dubbed x47.c, is offering specialised attack methods designed to drain the paid AI API credits of its victims. Researchers at Qrator Research Labs identified the malware, which is being advertised by a seller known as WraithTools.

One of the botnet’s primary features is an “AI API drain” command. This method uses valid API keys for services such as OpenAI, xAI, or compatible chat APIs to send repeated, billable requests directly to the provider. This type of attack is categorised by OWASP as a “denial of wallet” (DoW) attack.

Unlike traditional denial of service (DoS) attacks that target a website’s availability, a denial of wallet attack targets the victim’s finances. Because the malicious requests are sent directly to the AI provider using valid keys, the victim’s website or application can remain online while the underlying AI features continue to accrue significant costs. Qrator Research Labs noted that filtering traffic at the application level may not prevent these charges.

AI-Assisted Persistence and Stealth

The botnet also includes an “AI Stealth” module that leverages xAI’s Grok model. This module is designed to assess the infected host and select from various predefined actions to maintain persistence and avoid detection. If model calls fail, the malware is capable of using local fallbacks to continue its concealment efforts.

In addition to AI-driven capabilities, x47.c offers a range of traditional cybercrime tools. These include credential theft targeting browser passwords, cookies, and Discord tokens, as well as a SOCKS5 module that converts infected machines into relays for network traffic.

DDoS and Network Attacks

The malware provides 18 different attack methods, many of which focus on distributed denial of service (DDoS). These include:

• HTTP floods
• Slow HTTP connections
• TCP and UDP floods
• TLS connection stress
• Reflection and amplification techniques

The seller, WraithTools, has priced the botnet between $200 and $950, with higher tiers including advanced proxying and AI-assisted persistence features.

Mitigation Strategies

To defend against these threats, Qrator Research Labs recommends that organisations immediately revoke any exposed AI API keys. Security teams should monitor billing statements against legitimate usage patterns and implement strict spending limits on all AI service accounts. Disabling automatic top-up features can also prevent runaway costs during an attack.

Furthermore, organisations should perform endpoint cleanup to remove the malware and deploy DDoS protection at both the network and application layers.