Microsoft warns of large-scale AI-driven threat campaigns
Microsoft has today published a blog post regarding new threat campaigns in which the China-linked actor (APT) Storm-0940 is using the agentic AI framework ARTEX to automate attacks on a large scale.
T-Update
This live blog contains information about large-scale AI-driven threat campaigns
Update 21 October 2026 22:18
Microsoft has published a blog today regarding new threat campaigns in which the China-linked actor (APT) Storm-0940 uses the agentic AI framework ARTEX to automate attacks on a large scale. According to Microsoft, hundreds of organisations worldwide have been targeted, and in several cases, access to cloud environments with elevated privileges was obtained within a matter of hours.
According to Microsoft, the campaign uses, among other things:
- Abuse of compromised OAuth applications;
- Service principal and application identities;
- Microsoft Graph;
- Password spraying;
- Automated reconnaissance of cloud environments;
- Supply chain access via trusted third-party applications.
Microsoft has linked the campaign to a recent compromise of ConnectWise-related cloud applications. According to Microsoft, application identities were used to gain access to customers' Microsoft 365 environments. Successful authentication does not automatically mean that data has been accessed or exfiltrated, but Microsoft advises organisations to actively investigate their environments for potential indicators of compromise.
Background
Microsoft published a blog post today regarding new threat campaigns in which the China-linked actor (APT) Storm-0940 utilizes the agentic AI framework ARTEX to automate large-scale attacks. According to Microsoft, hundreds of organizations worldwide have been targeted, and in several instances, the threat actor gained access to cloud environments with elevated privileges within a matter of hours.
Risk
The campaign demonstrates that attackers are increasingly using AI to accelerate existing attack techniques and execute them at scale. The attack does not primarily target end users, but rather application identities, OAuth permissions, and trusted connections between vendors and cloud environments.
For organisations using Microsoft 365 and Microsoft Entra, it is therefore important to pay extra attention to:
- OAuth applications and granted permissions;
- Service principals;
- Application credentials and certificates;
- Unexpected Microsoft Graph activity;
- Changes to roles, permissions, and application configurations.
Advice
Based on the indicators, detection rules and hunting guidance published by Microsoft, Tesorion proactively conducts additional investigations within the environments of clients for whom we provide relevant monitoring and detection services.
In doing so, we look for, among other things:
- Known indicators from the Microsoft advisory;
- Suspicious service principal authentications;
- Anomalous OAuth activities;
- Unexpected changes to application permissions;
- Signs of potential privilege escalation or persistence.
If we identify relevant findings during these investigations, we will contact the client in question directly and provide additional guidance on how to proceed where necessary.
Microsoft has published extensive technical information, indicators and detection capabilities for organisations that wish to conduct their own investigations. Tesorion is actively monitoring developments regarding this campaign and will inform clients as new insights or additional risks become known. If you have any questions about this T-Update, please contact your regular point of contact or the Tesorion SOC.
Sources
Sign up to receive T-Updates
Receive the latest vulnerabilities in your email every Wednesday
More than 1,000 organizations have already joined us.
