Microsoft Warns State-Backed Threat Group Storm-0940 Uses Agentic AI in ConnectWise Supply Chain Compromise

Microsoft has revealed that a China-aligned state-sponsored threat group known as Storm-0940 was behind the recent, high-profile supply chain compromise targeting the ConnectWise Cloud Backup application. According to the technology giant’s latest threat intelligence disclosures, the adversary weaponized an advanced, Go-based agentic artificial intelligence penetration testing framework called ARTEX to execute high-speed network scanning, reconnaissance, password spraying, and downstream OAuth application abuses on a global scale.

The security incident underscores a troubling evolutionary leap in cyber warfare. By integrating agentic AI capabilities into their offensive operational workflows, threat actors can bypass the traditional human limitations of time, fatigue, and sequential processing. Microsoft reported that Storm-0940 utilized the ARTEX framework at "machine speed" to automate aggressive attack campaigns targeting hundreds of organizations worldwide since September 2026. The sweeping campaign has affected a diverse array of critical sectors, including aerospace and defense, automotive, chemicals, energy, financial services, government, legal, manufacturing, media, military, semiconductor, software, technology, think tanks, and non-profit organizations.

Security analysts emphasize that the integration of artificial intelligence fundamentally alters the calculus of modern cyber defense. Traditional security controls and monitoring systems are often calibrated to match the pace of human adversaries or standard automated scripts. However, agentic AI frameworks like ARTEX can operate autonomously around the clock, working continuously through weekends and holidays to pursue multiple breach paths simultaneously. This capability drastically compresses the timeline from initial targeting to achieving deep infrastructure penetration, making complex supply chain attacks significantly more devastating.

The mechanics of the recent campaign highlight how threat actors leverage established trust relationships within modern cloud ecosystems. Microsoft noted that Storm-0940’s use of ARTEX dramatically accelerated the speed at which compromised applications could be abused to authenticate into target environments. Because these applications had already been granted legitimate access to cloud resources, the threat actor was able to operate seamlessly without requiring interactive user sign-ins. Once authentication was achieved, the AI-accelerated follow-on activities immediately commenced, encompassing deep internal reconnaissance, application credential modifications, and calculated attempts to elevate privileges for expanded and persistent access.

Storm-0940 is far from a newcomer to the threat landscape. Microsoft first detailed the activities of this adversary in November 2024, shedding light on its utilization of a widespread botnet infrastructure known as Quad7, also tracked as CovertNetwork-1658, to execute highly evasive password-spraying attacks. Historical assessments indicate that the threat group has been actively engaged in targeted cyber operations since at least 2021. The adversary’s typical operational playbook involves exploiting vulnerable, internet-facing applications to gain initial footholds within corporate networks, subsequently leveraging compromised accounts to move laterally through targeted enterprise environments. Furthermore, the group has consistently demonstrated a capability to establish long-term persistence by appending unauthorized credentials to existing OAuth applications within cloud infrastructures, thereby maintaining continuous access to sensitive Microsoft 365 resources.

The transition toward automated, AI-driven operations has amplified the scale and velocity of these tactics. Microsoft’s telemetry indicates that a substantial portion of the recent ARTEX-driven attacks successfully yielded privileged access to organizational cloud infrastructures. In several documented instances, the entire operational chain—ranging from the initial targeting phase to the successful acquisition of privileged administrative access—unfolded in less than three hours.

The ConnectWise Cloud Backup incident provided a stark illustration of this unprecedented velocity. During the early stages of the operation, Storm-0940 deployed ARTEX to launch massive, automated sign-in attempts targeting more than 2,250 distinct cloud tenants, maintaining an astounding pace of 2,900 sign-in attempts per minute. Microsoft detailed that between October 10 and October 14, 2026, threat actor infrastructure hosted at a specific IP address conducted an automated cross-tenant password-spray campaign aimed at 327 individual accounts across eight different organizations. The primary objective was to secure unauthorized access to Microsoft 365 and Exchange Online environments.

During this intensive four-day operational window, Microsoft’s security monitoring systems captured over 151,000 failed logon events. The malicious activity reached a crescendo on October 11, recording a peak of 2,900 failures per minute and culminating in more than 111,000 authentication failures in a single calendar day. Telemetry showed that 48 distinct accounts were subjected to more than 1,000 failed attempts each. Despite the overwhelming volume of traffic, Microsoft noted that no successful authentications, token issuances, Graph API calls, or directory modifications were detected during this initial brute-force phase.

The operational dynamics shifted dramatically on October 15, 2026. On this date, the autonomous AI agents successfully acquired the identities of multiple legitimate ConnectWise cloud applications. Armed with these stolen app credentials, the agents began authenticating directly into customer environments, effectively exploiting the pre-existing trust relationships that these applications already held within victim tenants.

Following this successful authentication phase, the AI agents operating on behalf of Storm-0940 advanced well beyond basic access. Operating within Microsoft Graph and native directory operations, the automated agents initiated comprehensive internal reconnaissance. They executed targeted application credential modifications designed to establish robust persistence mechanisms, alongside unauthorized changes involving user identities, security groups, and privileged administrative roles.

Microsoft warned that the application credential modifications executed by Storm-0940’s automated agents within victim environments pose a long-term risk. These alterations can establish persistent backdoors through alternative identities, secondary credentials, and connected cloud services that remain fully functional and active even after the initial compromised application has been successfully mediated and isolated.

While Storm-0940 has become a prominent adopter of the framework, Microsoft’s global threat intelligence indicates that the weaponization of the ARTEX framework is not isolated to a single group. Cybersecurity researchers have observed multiple threat actors across diverse geographic regions integrating the tool into their offensive operational security workflows, signaling a broader industry trend toward the normalization of agentic AI in cybercrime.

The implications of these developments have drawn sharp responses from affected technology providers and the wider cybersecurity community. ConnectWise issued formal advisories addressing the incident, noting that the campaign reflects a sophisticated threat vector extending well beyond any single enterprise. The company emphasized that nation-state actors are increasingly focusing their efforts on the broader software supply chain because a single compromised, trusted connection can instantly open the door to hundreds or thousands of downstream organizations.

ConnectWise pointed out that Managed Service Providers occupy a uniquely vulnerable position in this interconnected chain, serving simultaneously as direct targets and as critical conduits into every client organization they support. Consequently, technology vendors and service providers alike must reevaluate their defense strategies with both risks continuously accounted for. In response to the breach, ConnectWise temporarily suspended several core services to contain the threat and initiated a phased restoration process designed to ensure system integrity, a recovery undertaking expected to require multiple days of meticulous validation before full operational capacity is restored.