---
title: "Google's AI Threat Tracker Report Reveals Rise in LLMjacking"
url: https://newscentral.site/google-s-ai-threat-tracker-report-reveals-rise-in-llmjacking/
language: en
publisher: "News Central Site"
section: "Technology"
published: 2026-11-16T14:00:00.000Z
updated: 2026-11-17T12:36:58.918Z
id: a8942b02-bbc1-42d3-b017-1e217573b7b4
source: "ZDNet https://www.zdnet.com/innovation/google-threat-intelligence-report-llmjacking-ai/"
attribution: "Link to https://newscentral.site/google-s-ai-threat-tracker-report-reveals-rise-in-llmjacking/ and name News Central Site when you quote or summarize this story."
---

# Google's AI Threat Tracker Report Reveals Rise in LLMjacking

Businesses are adopting AI and large language models (LLMs) rapidly, creating an opportunity for cybercriminals to exploit vulnerabilities through LLMjacking.

Businesses have rapidly adopted AI and large language models (LLMs) to streamline operations, from scheduling meetings to complex data analysis and coding tasks.

This has created a lucrative opportunity for cybercriminals, who are exploiting vulnerabilities through LLMjacking - the theft of AI-related credentials.

Google's Q4 AI Threat Tracker report revealed that gateways called LLM Mixers are emerging on the market, offering access to AI models for pennies on the dollar.

The cost of accessing these AI tools can vary significantly depending on the company. While some businesses spend over $10,000 per employee per month, others pay as little as $16.

However, it's not just the financial costs that are a concern - when employees have access to AI models, companies often introduce spending caps to prevent unauthorized usage.

But smaller outfits may overlook these limits, leaving their AI resources vulnerable to cybercriminals who can exploit them for malicious purposes.

The trend of LLMjacking is expected to continue into 2027 and beyond, according to analysts who have observed a significant increase in this type of cybercrime over the past year.

Unauthorized access to AI services has become a lucrative business, with illicit storefronts known as LLM Mixers providing users with access to these resources. These stores are often found on Russian- and Chinese-language marketplaces, where vendors offer various levels of access to AI models.

The types of models available for purchase vary widely, including frontier models from AI labs and jailbroken models marketed specifically for malicious purposes. Researchers have analyzed the advertisements offered by these vendors, finding that most provide a range of options for accessing AI resources.

Pricing for these services can be unpredictable, with some vendors charging relatively low prices while others demand much higher fees. In some cases, prices may increase significantly once certain thresholds are exceeded, such as when a prompt exceeds 100,000 tokens.

The cost of LLMjacking to businesses can be substantial, with estimates suggesting that it could exceed $46,000 per day on cutting-edge models. Some organizations have reported cumulative charges reaching over $1 million before access was revoked, highlighting the potential financial risks associated with this type of cybercrime.

While lower-tier services may seem affordable, the costs can quickly add up for businesses that rely heavily on AI resources. The potential consequences of LLMjacking make it essential for companies to prioritize cybersecurity and protect their AI assets from unauthorized access.

To mitigate LLMjacking risks, businesses must consider the infrastructure aspect of AI access being sold on the dark web. This involves a pool of stolen credentials and compute instances that route buyer traffic to ensure reliable access to enterprise AI models.

The sale of LLM infrastructure is often bundled with other services, creating complex attack chains. An LLM Mixer might offer access to an enterprise AI model, while its partners handle initial network breaches or purchase stolen credentials from malware operators.

To prevent LLMjacking and subsequent inclusion in a Mixer's catalog, organizations should focus on managing access, monitoring, and imposing restrictions. This is similar to securing SaaS applications and cloud setups against intrusions.

By setting usage limits and monitoring account activity, businesses can detect and respond to potential security incidents promptly. When usage spikes suddenly, alerts can be triggered to prevent massive bills from accumulating due to unauthorized AI service usage.

In order to minimize the risk of LLMjacking, organizations must prioritize cybersecurity measures that protect their AI assets from unauthorized access. This includes limiting token issuance, monitoring account activity, and imposing restrictions on AI service usage.

Organizations should take proactive measures to safeguard their AI access credentials from falling into the wrong hands. This involves securely storing and rotating API keys using a secrets manager, which helps prevent unauthorized access to sensitive information. Additionally, running regular audits can detect quiet probing, misconfigurations, or shadow accounts that may have gone unnoticed.

Implementing phishing protections and multi-factor authentication (MFA) is also crucial in making it more difficult for attackers to steal credentials or compromise accounts and networks. By taking these precautions, organizations can significantly reduce the risk of their enterprise AI access being compromised, thereby protecting their valuable assets.

---
Source: [ZDNet](https://www.zdnet.com/innovation/google-threat-intelligence-report-llmjacking-ai/)  
Published by News Central Site: https://newscentral.site/google-s-ai-threat-tracker-report-reveals-rise-in-llmjacking/
