Microsoft said on October 21 that it assesses a China-state aligned actor it tracks as Storm-0940 was responsible for the ConnectWise Cloud Backup supply-chain compromise, and that, after the ConnectWise credentials were compromised, the actor used ARTEX, an open-source agentic AI penetration testing framework, against downstream tenants. Microsoft does not say ARTEX was used to obtain the ConnectWise credential. This puts a name on the intrusion we described on October 19 in our earlier briefing.
What Microsoft reports
Microsoft describes Storm-0940 as active since 2021, with no known industry aliases, targeting multiple industries for data theft. It does not give a formal confidence level for the attribution. Microsoft says the activity did not result from a compromise of, or a vulnerability in, its own systems.
According to Microsoft, ARTEX runs LLM-driven agents that carry out a penetration test from reconnaissance to data exfiltration with little human involvement. It reports use as early as September 2026, and says several threat actors across regions are using the framework.
The reported sequence:
- October 10 to 14: a password-spray campaign from 144.172.64[.]99 against 327 accounts in eight organisations, with more than 151,000 failed logons. Microsoft says no successful authentication was observed in that campaign.
- After the ConnectWise access: sign-in attempts to more than 2,250 tenants using the compromised multi-tenant application credentials.
- Post-authentication: in some environments, Graph and directory activity including credential additions for persistence and role and group changes. Bulk secret reads across multiple Azure Key Vaults appear in Microsoft’s detection mapping, not its narrative.
Microsoft says several ARTEX-driven attacks gave the actor privileged access, including multiple instances where the time from initial targeting to privileged access was under three hours, and that “the majority of the targeting wasn’t successful.” It also cautions that an affected application or a successful sign-in does not by itself show data was accessed or stolen. Microsoft’s body text keeps the 2,250 and 2,900 figures apart, but its infographic puts them side by side and CyberScoop quotes Microsoft as joining them. Our reading: 2,900 per minute is the spray’s peak failure rate (body text), and 2,250 is the number of tenants with sign-in attempts. Microsoft’s graphic says “attempts per minute” where the text says “failures per minute”, and it does not define the pairing.
What it means for the sectors
Microsoft lists sectors targeted across the broader ARTEX campaign, including aerospace and defence, energy, financial services, government, manufacturing and semiconductors. These are not necessarily ConnectWise victims. In our assessment, finance and critical-infrastructure organisations that rely on MSPs for backup and IT administration inherit this exposure through the vendor, not through their own controls. The main change is speed: automation shortens the window between a vendor credential leaking and privileged access in your tenant.
Recommended actions
- Search every tenant for the ConnectWise application IDs, not display names. Microsoft publishes 506e08f9-6e9a-4da4-942b-a099a03644ae (Cloud Backup) among others.
- Keep affected ConnectWise connections disabled until ConnectWise issues formal restoration instructions, and preserve evidence before deleting enterprise applications.
- Review service-principal sign-ins from at least October 15, ideally from October 1, for 144.172.64[.]128, 144.172.64[.]99 and 144.172.64[.]60. Do not limit the search to published IPs.
- Audit credential additions, owner changes, role assignments, consent grants and Conditional Access changes. Our suggestion: check Key Vault access logs for bulk secret reads.
- Apply least privilege to multi-tenant application permissions and confirm vendor credential rotation closed remaining access.
Attribution here rests on Microsoft’s assessment alone. We found no independent confirmation of the Storm-0940 link as of October 23, 2026, but our only sources are Microsoft and CyberScoop, which repeats Microsoft, so that is weak evidence. We rate severity “high” as our own assessment; most targeting reportedly failed.