Google's Threat Intelligence Group reported in its Q4 AI Threat Tracker that criminals are now using gateways called 'LLM Mixers' to resell stolen enterprise AI credentials and tokens, sometimes for as little as $0.10 per batch. Victim organizations can be left with usage bills exceeding $100,000 after attackers exploit stolen API keys or logins to run unauthorized AI workloads. GTIG's chief analyst said LLMjacking incidents rose sharply through 2026.
zdnet.com
· 2026-11-16
Google's Threat Intelligence Group reports that state-sponsored and financially motivated hackers are increasingly using open-source AI pentesting frameworks—including CAIRN, ARTEX, CyberStrikeAI and HexStrike AI—to automate attack chains. Examples cited include CAIRN deployed against a U.S. aerospace target, suspected North Korean actors using Claude Code to deliver malware, and a multi-agent tool called 'exploit_agent' used against French media and education sites. The FBI separately warned last week that the ARTEX framework has been weaponized against financial, government, defense and healthcare organizations since August.
bleepingcomputer.com
· 2026-11-16
Google's Threat Intelligence Group dismantled x.cpe, a botnet that infected over 1.3 million Google-certified Android TV boxes, mostly sold under the MECOOL brand by Shenzhen-based Videostrong Technology. The malware used a backdoor to remotely install itself after devices went online, disabling Play Protect and evolving into a three-stage platform sold to third parties for residential proxying and ad fraud.
arstechnica.com
· 2026-11-03
Google Threat Intelligence Group and Mandiant report that cybercriminals are stealing access to commercial large language models through compromised cloud credentials and reselling that access on underground markets. In some cases, fraudulent usage charges topped $100,000 in a single day, with total costs to a victim organization reaching $1 million before the compromised access was cut off.
darkreading.com
· 2026-11-03
Google's Threat Intelligence Group and Mandiant found that attackers are stealing API keys and exploiting exposed LLM endpoints to resell unauthorized access through services called LLM mixers. Victim organizations have seen fraudulent usage spike past 1,000 requests per second, with some racking up over $100,000 in charges in a single day and cumulative losses reaching $1 million. One service, Jet Night, reportedly passed nearly $1 million in token costs to a single victim after executing more than 500,000 API calls in under eight hours.
bleepingcomputer.com
· 2026-11-03
A coalition including Google Threat Intelligence Group, TrendAI and Japan's National Police Agency disrupted a botnet of more than 1.3 million Android TV boxes and projectors spanning over 200 countries. Google says Shenzhen-based manufacturer Videostrong Technology deliberately built a backdoor called XCPE into device firmware and sold access to criminal affiliates, who installed malware modules to route traffic through unsuspecting owners' home networks. The group also took down command-and-control domains tied to affiliate modules and said it has cut ties with Videostrong, banning it from shipping devices with preloaded Google apps.
it.slashdot.org
· 2026-11-03
Google's Threat Intelligence Group worked with industry partners to take down a botnet, dubbed XCPE, built from more than 1.3 million Android TV set-top boxes. Investigators traced the malware to Videostrong Technology Co., a Chinese manufacturer accused of preloading devices with backdoors that resold users' internet connections through residential proxy services and enabled advertising fraud. Google took legal action to seize the domains and servers used to control the botnet.
krebsonsecurity.com
· 2026-11-03
Google Threat Intelligence Group analyst John Hultquist told the Financial Times that his team has observed a major increase in LLMjacking throughout 2026. The technique involves criminals stealing API keys or login credentials to hijack businesses' AI accounts, exploiting high or unlimited usage limits to run up token costs or misuse compute resources without paying.
zdnet.com
· 2026-09-29
A Google Threat Intelligence Group analyst named Larsen ran a covert operation to observe the hacker collective TeamPCP, gaining access to a server where the group stored stolen credentials from numerous victim companies. Google used this visibility to rapidly notify cloud providers like AWS and Microsoft to revoke compromised access tokens before the hackers could exploit them for extortion, rather than trying to contact each victim individually.
arstechnica.com
· 2026-09-20
Google's Threat Intelligence Group disclosed a credential-harvesting operation where attackers compromised cloud infrastructure and deployed a multi-agent AI framework that scanned for vulnerabilities, fixed its own errors, and rotated IP addresses largely without human input. The entire campaign, which compromised thousands of third-party credentials, took less than six hours to execute. Separately, Microsoft found AI-assisted phishing messages reaching click-through rates of 54%, compared to roughly 12% for conventional campaigns.
bleepingcomputer.com
· 2026-09-17
Google's Threat Intelligence Group reports that attackers are moving from simple AI-assisted coding to fully autonomous, multi-agent systems that plan and execute entire attacks with little human input. In one case, a financially motivated actor compromised cloud infrastructure and used an AI coding chatbot with markdown-based agent instructions to build and run a mass credential-harvesting operation in under six hours. Separately, researchers found an exposed command-and-control server running a framework called 'Recon' that autonomously managed reconnaissance and tens of thousands of harvested credentials.
bleepingcomputer.com
· 2026-09-08