Storm-0940 Exploits AI Framework to Breach 2,250 Microsoft Tenants via ConnectWise

Microsoft has identified a sophisticated supply chain attack that leveraged artificial intelligence to compromise more than 2,250 Microsoft 365 tenants. The campaign was attributed to Storm-0940, a threat actor aligned with the Chinese state. The group utilized a stolen set of credentials from ConnectWise, a multi-tenant application, to bypass standard user authentication. In some instances, the attackers achieved privileged access within three hours of initial targeting, highlighting the speed of the intrusion.

The operation marks a significant shift in cyber tactics by employing ARTEX, an open-source agentic AI framework. Microsoft describes the system as a collection of large language model agents that conduct autonomous penetration tests. This technology accelerated the scanning, reconnaissance, and password spraying processes. By operating continuously with minimal human intervention, the agents left defenders with a narrow window to detect and respond to the unfolding breach.

Storm-0940 has been active since 2021 and targets specific sectors including aerospace, defense, semiconductors, and government entities. Microsoft clarified that the attack did not originate from a vulnerability in its own infrastructure but rather from compromised third-party application credentials. A parallel password-spraying campaign attempted over 151,000 logons across eight organizations, though most attempts failed and were locked out. The most effective vector involved using four specific application credentials to sign into customer tenants from a single IP range.

Security firm Huntress observed the attackers moving through tenants in a sorted order with seconds between actions, suggesting the use of automated scripts. On October 15, one tenant recorded an app assigning itself the Global Administrator role. By October 16, unauthorized sign-ins were confirmed at 162 tenants. The following day, nearly 1,000 tenants showed activity from the attacking IP address, with more than 300 sign-ins succeeding. The attackers also planted new secrets configured to expire in 2028, ensuring long-term access.

The use of ARTEX distinguishes this campaign from traditional attacks, as the framework can integrate with various AI models. Public reports this month linked the tool to breaches at South Korean financial institutions. The developer of ARTEX, identified as Autumn-27, announced on GitHub that the project would stop receiving updates and move to closed source. This development may impact other threat actors who rely on the open-source component for their operations.

ConnectWise responded by suspending several services, including Cloud Backup and SaaS Security, to contain the damage. The company has initiated a phased restoration process that is expected to take several days. Preliminary checks indicate no evidence that partner backup data was compromised. Microsoft advised customers to keep affected app connections disabled until formal restoration instructions are provided. Organizations are urged to review their logs from this month to identify any anomalous activity associated with the breach.

As the restoration efforts continue, the incident underscores the growing threat posed by AI-driven automation in cyberattacks. The rapid movement through tenants and the use of advanced tools like ARTEX suggest a high level of sophistication. Companies relying on multi-tenant SaaS applications should monitor their authentication logs closely for signs of unauthorized access or privilege escalation in the coming days.