CrowdStrike said on 29 October 2026 that a China-nexus hacking group it tracks as VAPOR PANDA used an AI agent framework called ARTEX, “alongside the Chinese-developed DeepSeek large language model (LLM) and a second commercial LLM”, to run intrusions against “more than 200 distinct organizations worldwide” beginning on 17 October.
NBC News, citing CrowdStrike, reported that the campaign “illustrates how AI is allowing hackers to operate on a scale that previously required far more human effort”. By our arithmetic, more than 200 organizations in an 11-day window is about 18 a day.
Short explainer · China-linked hackers ran DeepSeek agents
A vertical news Short narrated from this piece’s headline and opening line.
Business Pill · Least Privilege
A one-minute explainer of least privilege: giving each system or person only the access a task needs. It teaches the general idea only and says nothing about any company or group in this story.
The key insight: By CrowdStrike’s account, a single ARTEX controller directed AI agents running mainly on DeepSeek against more than 200 organizations in 11 days, with a human operator steering them at decision points; by our arithmetic that is about 18 targets a day, with confirmed post-exploitation at under 5% of them.
CrowdStrike said it observed exploitation activity from a single ARTEX controller “directed against entities in at least 14 countries over an 11-day window”. Targets included “foreign ministries in Latin America, Europe, Asia, and the Middle East”, U.S. semiconductor, defense and think tank organizations, and “telecommunications providers in multiple regions”.
One engagement “targeted 19 U.S.-based organizations in the pharmaceutical, semiconductor, defense, nuclear, and financial sectors”, CrowdStrike said. It “confirmed post-exploitation activity at nine organizations” and said the group also ran password-spraying and credential-scanning activity against “at least 12 additional organizations”. By our arithmetic, nine confirmed intrusions are under 5% of the more than 200 organizations targeted.
VAPOR PANDA’s ARTEX campaign from 17 October 2026, per CrowdStrike: more than 200 organizations targeted over 11 days in at least 14 countries, password-spraying or credential scanning against at least 12 more, and confirmed post-exploitation at nine. 200+ in 11 days = about 18 a day; 9 of 200+ = under 5%, our arithmetic. Source: CrowdStrike blog, 29 Oct 2026.
How the Agents Worked
According to CrowdStrike, VAPOR PANDA “ran ARTEX primarily with the Chinese-origin DeepSeek LLM, supplemented with a second, subscription-based commercial LLM that CrowdStrike Intelligence could not confirm”. In a single intrusion, it said, the planning agents “consumed on the order of 1.4 billion input tokens”.
ARTEX “coordinates a main agent with subordinate planner and worker agents over a shared graph of facts, findings, and goals”, the report said, and “kept a human in the loop”: recovered transcripts showed an operator issuing instructions in Chinese. CrowdStrike said the operator also directed the agents to find Windows machines without its Falcon sensor.
CrowdStrike called the campaign “an escalation in China-nexus use of AI agents, from single intrusions to multiple simultaneous engagements directed by a human operator”. It said the activity is separate from the ARTEX activity reported by industry sources under the Storm-0940 designation.
What CrowdStrike Expects
CrowdStrike assessed that “China-nexus adversaries will highly likely continue adopting agentic frameworks such as ARTEX over the next 12 months”, pairing them with commercially available LLMs. It said the group’s reliance on DeepSeek and other commercial models “likely reflects a deliberate effort to construct a resilient toolchain from multiple model providers rather than depend on any single service”.
The Structural Read
As we read CrowdStrike’s report, the change it describes is in scale rather than technique: the agents chained familiar methods, and CrowdStrike says this breadth likely enables a single operator to apply a wider range of techniques against more targets at once.
As we read it, the model layer was interchangeable: CrowdStrike says the group ran DeepSeek alongside a second commercial model it could not identify, through a purpose-built harness.
By our arithmetic, more than 200 organizations in 11 days is about 18 a day, and nine confirmed post-exploitation cases are under 5% of them.
CrowdStrike, 29 Oct 2026
“This campaign marks an escalation in China-nexus use of AI agents, from single intrusions to multiple simultaneous engagements directed by a human operator”
Three Implications
FOR MODEL PROVIDERS CrowdStrike says the group ran ARTEX primarily with DeepSeek, supplemented with a second, subscription-based commercial model it could not confirm.
FOR DEFENDERS CrowdStrike says the operator directed the agents to find Windows machines without its Falcon sensor, building sensor checks into the workflow.
FOR THE NEXT 12 MONTHS CrowdStrike assessed that China-nexus adversaries will highly likely continue adopting agentic frameworks such as ARTEX over the next 12 months.
The framework says: “The agentic surface — the harness — is becoming where value pools in the entire AI stack.”
As we read it, CrowdStrike’s report describes the same layer from the attack side: a purpose-built harness coordinating agents across models. The framework was published by The Business Engineer and does not discuss CrowdStrike, VAPOR PANDA or ARTEX.
The figures are CrowdStrike’s, from its own telemetry and an exposed controller; CrowdStrike says the 1.4 billion tokens covered a single intrusion and did not give an operation-wide total. It could not confirm which second commercial model was used. We read CrowdStrike’s blog post and the opening of NBC’s report; we did not contact DeepSeek, the targeted organizations or the Chinese government.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
CrowdStrike says a China-nexus group used DeepSeek-driven ARTEX agents against more than 200 organizations in 11 days, confirming post-exploitation at nine. By our arithmetic, that is about 18 targets a day, with a single intrusion consuming about 1.4 billion input tokens, according to CrowdStrike.
A note on sourcing. Facts come from CrowdStrike’s blog post of 29 October 2026 and the opening of NBC News’ report of the same day, both read on 29 October. We did not contact CrowdStrike, DeepSeek or the targeted organizations. Figures marked as ours are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
