AI-driven, large-scale nation-state campaign using ARTEX enables supply chain compromise

* Storm-0940, a China-state aligned actor, used ARTEX starting September 2026 to automate attacks globally.
* Attacks targeted hundreds of organizations across numerous industries.
* Some ARTEX-driven attacks resulted in privileged access to cloud infrastructure within three hours.
* Storm-0940 was responsible for the ConnectWise supply chain compromise.
* ARTEX accelerated scanning, reconnaissance, password spraying, and OAuth application abuse during the compromise.
* Activity involved automating authentication to Microsoft 365 customer environments without interactive sign-in.
* AI-accelerated activity was observed through Microsoft Graph, including reconnaissance and privilege modification attempts.
* Automated activities generated high volumes of traffic and log data.
* A password-spray campaign targeted 327 accounts across eight organizations between October 10 and 14, 2026.
* This campaign generated over 151,000 failed logon events.
* Follow-on activity involved agents using illicit access to authenticate and perform Graph operations in customer environments.
* Indicators of compromise include IP addresses such as 144.172.64[.]99, 144.172.64[.]60, and 144.172.64[.]128.

The narrative demonstrates a significant shift in the kinetic capabilities of threat actors, moving from manual exploitation to leveraging agentic AI frameworks to achieve operational scale and speed. The core implication is that the integration of autonomous agents drastically reduces the time-to-objective for complex supply chain attacks, transforming reconnaissance and credential abuse into machine-speed operations that evade traditional human-centric defensive response times. The persistence achieved through manipulating service principal sign-ins highlights a critical vulnerability in cloud identity management where application trust can be weaponized beyond their intended scope. The tension lies between the observed technical reality—the sheer speed of AI-accelerated activity—and the necessary human capacity for effective containment and long-term security governance. The focus shifts from merely blocking initial access to monitoring and controlling the complex, autonomous interactions occurring within identity and API layers where persistence is established.
Bridge Questions: If agentic systems can accelerate compromise exponentially, what non-linear defensive strategies must organizations adopt to effectively counter operations that operate at machine speed? How should risk models be updated to account for the potential reach of AI-driven automation when assessing third-party application trust versus user intent? What are the long-term governance implications when privileged access is established via automated application credential modification rather than traditional human compromise?