← Voltar ao feed
medium

Google GTIG: 'LLM Mixers' resell stolen enterprise AI credentials and proxy access, driving LLMjacking bills above $100,000

1 fonte(s) · 1 atualização(ões) · publicado 16/11/2026 11:00 · coletado há 1 dia

Google Threat Intelligence Group's Q4 AI Threat Tracker describes LLM Mixers, underground storefronts that resell unauthorized access to enterprise AI services and the proxy infrastructure behind them. GTIG analyzed LLMjacking vendor ads on Russian- and Chinese-language marketplaces. Most vendors offer frontier models and jailbroken models marketed for malicious use. Token prices ranged from $0.10 to $25, and one vendor, APIMart, raised its price fivefold above 100,000 tokens. Victims pay the usage bills: Sysdig estimates costs of roughly $46,000 on earlier models and over $100,000 per day on cutting-edge ones, and Mandiant recorded cases above $100,000 in 24 hours, with cumulative charges reaching $1 million at some organizations. The article is a news report and cites no CVE, IOC or named threat actor. No primary Google report was opened for this record.

Detalhes técnicos

  • LLMjacking is the theft of usernames, passwords or API keys for AI services. Infostealer operators and initial access brokers can feed this supply chain.
  • Mixers route buyer traffic through a pool of stolen credentials and compute instances. They spread query volume across multiple victim environments to reduce the chance of detection.
  • Misuse includes running high-compute tasks, operating malicious AI deployments, extracting data from AI memory and poisoning training data.

Remediação

  • Set token and spending caps on every AI service and account, and alert on usage spikes.
  • Store API keys in a secrets manager and rotate them frequently.
  • Audit for quiet probing, misconfigurations and shadow accounts.
  • Deploy phishing protection and MFA.

Linha do tempo

  1. zdnet.com ↗ · pub. 16/11/2026 11:00 · coletado 16/11 12:00 primeiro registro

    Google GTIG: 'LLM Mixers' resell stolen enterprise AI credentials and proxy access, driving LLMjacking bills above $100,000

    ler a análise desta fonte

    Google Threat Intelligence Group's Q4 AI Threat Tracker describes LLM Mixers, underground storefronts that resell unauthorized access to enterprise AI services and the proxy infrastructure behind them. GTIG analyzed LLMjacking vendor ads on Russian- and Chinese-language marketplaces. Most vendors offer frontier models and jailbroken models marketed for malicious use. Token prices ranged from $0.10 to $25, and one vendor, APIMart, raised its price fivefold above 100,000 tokens. Victims pay the usage bills: Sysdig estimates costs of roughly $46,000 on earlier models and over $100,000 per day on cutting-edge ones, and Mandiant recorded cases above $100,000 in 24 hours, with cumulative charges reaching $1 million at some organizations. The article is a news report and cites no CVE, IOC or named threat actor. No primary Google report was opened for this record.

    Technical Details

    • LLMjacking is the theft of usernames, passwords or API keys for AI services. Infostealer operators and initial access brokers can feed this supply chain.
    • Mixers route buyer traffic through a pool of stolen credentials and compute instances. They spread query volume across multiple victim environments to reduce the chance of detection.
    • Misuse includes running high-compute tasks, operating malicious AI deployments, extracting data from AI memory and poisoning training data.

    Remediation

    • Set token and spending caps on every AI service and account, and alert on usage spikes.
    • Store API keys in a secrets manager and rotate them frequently.
    • Audit for quiet probing, misconfigurations and shadow accounts.
    • Deploy phishing protection and MFA.