LLMjacking 2026: Stolen API Keys Are Costing Enterprises Up to $1 Million in Cloud AI Charges
LLMjacking 2026: Stolen API Keys Are Costing Enterprises Up to $1 Million in Cloud AI Charges
In 2026, Google Threat Intelligence Group and Mandiant documented a surge in LLMjacking attacks targeting enterprises across financial services, retail, technology, and transportation sectors in North America, South America, and Asia Pacific. Threat actors—including the tracked cluster UNC6993—obtained cloud credentials via phishing, misconfigured Service Account keys, exposed GitHub Personal Access Tokens, and public code repositories to gain unauthorized access to enterprise LLM APIs. After establishing a foothold, attackers escalated privileges, performed multi-region reconnaissance, cloned codebases for hard-coded secrets, and created backdoor service accounts. Fraudulent LLM consumption exceeded $100,000 USD in a single day at individual organizations, with cumulative charges reaching $1 million USD before access was revoked.
LLMjacking represents a structural evolution beyond cryptojacking, exploiting the high per-token cost of frontier AI models and the emergence of underground LLM Mixer services that proxy and resell stolen enterprise LLM access. As agentic AI workflows proliferate in production cloud environments, the attack surface for credential-based API abuse is expanding rapidly, making cloud identity hygiene and real-time egress enforcement urgent priorities for security teams in 2026 and beyond.
High impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(high)
Vapor Panda's ARTEX AI Campaign: How China-Nexus Hackers Automated Attacks on 200+ Organizations
Beginning October 17, 2026, a China-nexus threat actor tracked as Vapor Panda deployed ARTEX, an open-source multi-agent AI penetration-testing framework, in a sweeping cyberespionage campaign targeting more than 200 organizations across 14 countries. Victims span telecommunications providers in the United States and Belgium, government ministries in Greece, the Middle East, and Latin America, pharmaceutical and semiconductor firms, defense contractors, and financial institutions. AI agents orchestrated by DeepSeek and a second commercial LLM autonomously performed reconnaissance, exploited vulnerabilities, deployed web shells, harvested credentials, and pivoted laterally—all while human operators issued instructions in Chinese and intervened only at critical decision points. At least nine organizations were fully compromised, and the BLOODALCHEMY backdoor was deployed against government targets in Latin America and the Middle East.
This campaign represents a structural shift in nation-state tradecraft: AI agents now enable a single operator team to run dozens of simultaneous intrusions at machine speed, dramatically lowering the cost of large-scale cyberespionage. The DIVD breach—where ARTEX chained two zero-days to achieve root in 12 minutes—illustrates how agentic AI compresses the attacker's kill chain to intervals too short for human-speed defenses to interrupt.
Critical impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(medium)
EExfiltration(high)
IImpact(medium)
SysGoRAT and AI-Generated Scripts: How CL-UNK-1211 Weaponized LLMs Against Government Linux Infrastructure
Between October 2025 and April 2026, a threat cluster designated CL-UNK-1211 conducted a sustained cyber espionage campaign against government Linux server infrastructure. Attackers deployed SysGoRAT, a custom Go-based remote access trojan communicating over WebSocket C2, alongside AI-generated PHP and shell scripts for reconnaissance, credential harvesting, and email exfiltration. The group replaced .bashrc files, injected code into PHP login pages, and scanned configuration files for cleartext credentials. Attempts to compress and exfiltrate government mailboxes and social media account data were observed, alongside deliberate log wiping and deployment of red-team markers to disguise malicious activity as legitimate penetration testing.
This campaign reflects an accelerating trend of state-affiliated threat actors integrating large language models and agentic AI frameworks to compress attack timelines, reduce manual overhead, and rapidly adapt post-exploitation tooling. As agentic offensive frameworks like ARTEX, PentAGI, and Hexstrike proliferate, defenders must reckon with machine-speed intrusion cycles that outpace conventional detection and response cadences.
Critical impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(high)
Vapor Panda Used Agentic AI to Compromise 140 Targets in 11 Days—Here Is What That Means for Your Cloud Security
Beginning October 17, a China state-sponsored threat actor tracked by CrowdStrike as Vapor Panda deployed ARTEX—an agentic AI-powered penetration-testing framework paired with DeepSeek—to compromise approximately 140 of more than 200 targeted organizations across 14 countries in just 11 days. Targets spanned foreign ministries, defense and intelligence agencies, a European head-of-government office, and 19 U.S. organizations in pharmaceuticals, semiconductors, defense, nuclear, and financial sectors, as well as telecom providers in the United States, Belgium, and the Philippines. ARTEX enabled attack commands at rates exceeding 1,500 per hour—roughly 37 times the pace of traditional hands-on-keyboard intrusions—allowing simultaneous multi-target exploitation at near-zero marginal cost per victim.
This campaign marks the first confirmed nation-state deployment of agentic AI at scale for offensive cyber operations. CrowdStrike reports agentic adversary attacks grew nearly threefold month-over-month, from roughly 45 incidents in September to 126 in October 2024 and climbing, signaling a structural shift in the speed, breadth, and economics of state-sponsored intrusion campaigns that defenders must urgently address.
Critical impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(medium)
LMLateral Movement(medium)
C&CCommand & Control(medium)
EExfiltration(medium)
IImpact(high)
AI-Powered ARTEX: How Vapor Panda Turned Agentic LLMs Into a Weapon Against 200+ Global Organizations
Beginning October 17, 2026, the Chinese state-sponsored threat actor Vapor Panda (also tracked as Earth Estries) weaponized ARTEX, a Go-based agentic AI penetration-testing framework, to compromise or probe more than 200 organizations across foreign ministries, semiconductor firms, defense contractors, telecommunications providers, and financial institutions worldwide. Leveraging DeepSeek and a commercial LLM, ARTEX autonomously coordinated planner and worker agents to execute reconnaissance, credential spraying, SAML/XXE chaining, container escapes, and lateral movement—consuming roughly 1.4 billion input tokens in a single intrusion. Post-exploitation was confirmed at nine organizations, where the group deployed BLOODALCHEMY, a ShadowPad derivative, enabling persistent covert access at scale previously unachievable through manual operations.
This campaign marks a qualitative shift in state-sponsored cyber espionage: agentic AI now allows a small operator team to run concurrent, multi-path intrusions across dozens of targets simultaneously. As ARTEX proliferates among multiple threat clusters, defenders must urgently reassess detection strategies, east-west segmentation, and egress controls designed for human-paced adversaries, not autonomous AI agents operating at machine speed.
High impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(medium)
ESET APT Report Q2–Q3 2026: How China, Iran, North Korea, and Russia Are Reshaping the Threat Landscape
Between April and September 2026, ESET researchers documented an expansive wave of APT activity spanning China-, Iran-, North Korea-, Russia-, and Belarus-aligned threat actors. China-aligned groups pursued semiconductor espionage in Taiwan and fraud-linked intelligence collection across Southeast Asia and Oceania. Iran-aligned actors including FilthyFeline and MuddyWater resumed operations against Israeli and US targets after a period of reduced activity. North Korea's Lazarus Group expanded into drone-sector espionage tied to the Russia-Ukraine conflict, while DeceptiveDevelopment pivoted to software supply-chain compromises to harvest developer credentials. Russia-aligned groups continued targeting European defense and government entities, with Belarus-aligned FrostyNeighbor executing the first documented destructive cyberattack by a Belarus-linked actor in the European theater, striking a Ukrainian transportation manufacturer.
This reporting period marks a meaningful escalation in APT operational breadth and destructive ambition. The convergence of Russia-Belarus cyber-proxy activity, North Korea's drone-sector targeting, and the widespread adoption of AI-assisted malware development signals a new phase in state-sponsored threat evolution that defenders across critical infrastructure, semiconductors, and financial services must urgently address.
Medium impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(medium)
LMLateral Movement(medium)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(high)
Advantest Ransomware Attack Exposes Critical Gaps in Semiconductor Supply Chain Security
In February 2026, Advantest Corporation, a leading Japanese semiconductor test equipment manufacturer, suffered a ransomware attack that compromised sensitive personal information of individuals. Threat actors gained unauthorized access to company systems, deployed ransomware, and exfiltrated data containing contact information, Social Security numbers, driver's licenses, passport numbers, medical information, and financial data. The company took eight months to confirm the data breach, notifying affected individuals in October 2026 and offering 18-month identity monitoring services through Kroll.
This incident highlights the continuing evolution of ransomware groups targeting critical semiconductor supply chain companies, exploiting the industry's increasing digitization and interconnected manufacturing processes that are vital to global technology infrastructure.
High impact
Kill Chain
ICInitial Compromise(low)
PEPrivilege Escalation(medium)
LMLateral Movement(medium)
C&CCommand & Control(medium)
EExfiltration(high)
IImpact(high)
State-Sponsored Cyber Attacks on AI Supply Chain in 2026
In 2026, the global race to dominate artificial intelligence (AI) has intensified, with nations vying for control over critical minerals, semiconductor production, and AI model development. This competition has led to increased state-sponsored cyber operations targeting every link in the AI supply chain, from mining companies to data centers and AI research institutions. Notably, Chinese state-sponsored hackers have been implicated in sophisticated cyber espionage campaigns aimed at extracting sensitive information and disrupting competitors' advancements in AI technologies.
The urgency of securing the AI development chain has never been more critical. As AI becomes deeply integrated into various sectors, the potential for cyber threats to disrupt economies and national security has escalated. Organizations must adopt comprehensive cybersecurity strategies to protect against these evolving threats, ensuring the resilience of their AI infrastructures.
Medium impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(high)
Analog Devices 2026 Data Breach: ExfilSquad's Latest Target
In June 2026, Analog Devices, a leading semiconductor company, detected unauthorized access to certain company systems, resulting in the exfiltration of unspecified files. The company promptly activated its incident response protocols and engaged external cybersecurity experts to contain the breach. As of now, there is no evidence that the stolen data has been leaked online or used for fraudulent purposes. Business operations remain unaffected, and the company does not anticipate any material impact on its financial condition.
This incident underscores the growing threat posed by data extortion groups like ExfilSquad, which claimed responsibility for the breach. Organizations must remain vigilant and enhance their cybersecurity measures to protect sensitive information from such emerging threats.
Medium impact
Kill Chain
ICInitial Compromise(low)
PEPrivilege Escalation(low)
LMLateral Movement(low)
C&CCommand & Control(low)
EExfiltration(low)
IImpact(low)
Unpatchable Vulnerability in Tangem Wallets Exposed by Laser Attack
In July 2026, Ledger's Donjon security team disclosed a vulnerability in Tangem crypto wallet cards, revealing that a precisely timed laser pulse aimed at the card's secure element chip can reset the card's password without the original password or backup card. This allows an attacker to gain control over the wallet and transfer funds. The attack requires physical possession of the card, specialized equipment estimated at $250,000, and leaves visible damage, making it impractical for widespread exploitation. However, due to Tangem's design, which lacks firmware update capabilities, this vulnerability cannot be patched, leaving all existing cards susceptible.
This incident underscores the challenges in securing hardware wallets against sophisticated physical attacks and highlights the importance of considering firmware update mechanisms in device design. While the attack's complexity limits its immediate threat, it raises concerns about the long-term security of devices that cannot receive updates to address discovered vulnerabilities.
Low impact
Kill Chain
ICInitial Compromise(high)
PEPrivilege Escalation(high)
LMLateral Movement(high)
C&CCommand & Control(high)
EExfiltration(high)
IImpact(high)
Tata Electronics Cyberattack: A Wake-Up Call for Supply Chain Security
In June 2026, Tata Electronics, a division of the Tata Group specializing in electronic components and semiconductor manufacturing, confirmed a cyberattack that impacted parts of its IT infrastructure. The company stated that operations remained unaffected. The World Leaks threat group claimed responsibility, leaking over 200,000 files totaling approximately 630 GB, including sensitive manufacturing data for Apple and Tesla products. The leaked information comprises internal component schematics, PCB designs, material specifications, and SDK files. ([business-standard.com](https://www.business-standard.com/companies/news/tata-electronics-hit-by-cyber-breach-exposing-apple-tesla-trade-secrets-126062201241_1.html?utm_source=openai))
This incident underscores the escalating threat posed by data extortion groups like World Leaks, which focus on stealing and leaking sensitive corporate data without deploying traditional ransomware. The breach highlights the critical need for robust cybersecurity measures and supply chain security, especially for companies handling proprietary information of major technology firms. ([business-standard.com](https://www.business-standard.com/companies/news/tata-electronics-cyber-breach-apple-tesla-supply-chain-security-126062300396_1.html?utm_source=openai))
High impact
Kill Chain
ICInitial Compromise(medium)
PEPrivilege Escalation(medium)
LMLateral Movement(medium)
C&CCommand & Control(medium)
EExfiltration(high)
IImpact(high)
Foxconn's 2026 Ransomware Breach: A Wake-Up Call for Manufacturing Cybersecurity
In May 2026, Foxconn, a leading electronics manufacturer, confirmed a cyberattack affecting several of its North American facilities. The Nitrogen ransomware group claimed responsibility, alleging the theft of 8 terabytes of data, including confidential project files from major clients such as Apple, Nvidia, Intel, Google, and Dell. The attack disrupted operations, forcing some employees to revert to manual processes or halt work temporarily. Foxconn's cybersecurity team responded promptly, implementing measures to restore normal production.
This incident underscores the escalating threat to the manufacturing sector, which has seen a significant rise in ransomware attacks due to its critical role in global supply chains and low tolerance for operational downtime. The breach highlights the need for robust cybersecurity measures to protect sensitive data and maintain business continuity.
High impact
Kill Chain
ICInitial Compromise(medium)
PEPrivilege Escalation(medium)
LMLateral Movement(medium)
C&CCommand & Control(medium)
EExfiltration(high)
IImpact(high)
