Skip to content
Microsoft Ties Chinese-Linked Hackers to AI-Powered Breach at ConnectWise, Hundreds ...

Microsoft Ties Chinese-Linked Hackers to AI-Powered Breach at ConnectWise, Hundreds ...

Finance.Biggo • October 22, 2026

Microsoft disclosed Wednesday that a hacking group it tracks as Storm-0940 used an AI-driven penetration testing framework to breach IT services provider ConnectWise, part of a broader campaign that touched hundreds of organizations around the world.

The group, which Microsoft describes as aligned with the Chinese state, carried out the intrusion using ARTEX, an open-source tool that lets operators run any AI model against automated penetration testing tasks. According to the software giant, ARTEX can be configured to execute autonomous jobs, including reconnaissance and data exfiltration.

ConnectWise CEO Manny Rivelo confirmed the attribution in a separate statement, saying that based on Microsoft's analysis and evidence available to his company, "it appears that a China state-aligned group was responsible for the attack on ConnectWise Cloud Backup." He said the company had "contained the affected application connections and are restoring services in controlled phases," while cautioning that those steps "do not remove unauthorized access that may have already been established within individual customer tenants, which require separate review."

ConnectWise had already acknowledged the incident on October 16, when a notice on its website said it was "investigating recent attempts of unauthorized access associated with the ConnectWise Cloud Backup environment." A company spokesperson did not immediately respond to a request for on Microsoft's conclusions.

According to Microsoft, the intruders rode the trusted identity of the ConnectWise application to pass authentication without an interactive login. Because ConnectWise runs IT operations on behalf of managed service providers, that foothold potentially opened a path into many customer environments at once rather than requiring the attackers to break into each business individually.

The targeting spanned an unusually broad set of industries, including aerospace, defense, automotive, chemicals, energy, financial services, government, legal, manufacturing, media, military, semiconductor, software, technology, think tanks and nonprofits.

Most of the attempts failed, Microsoft said, but Storm-0940 still managed to establish "privileged cloud persistence" - a level of access that survives cleanup efforts - at numerous organizations within a short window. In some cases, the group obtained privileged access within three hours of first targeting a victim, according to the company.

Microsoft also pointed to reports linking ARTEX to earlier attacks, including a data leak that hit South Korean financial institutions. More than 16 financial companies in the country have been recorded as losing upward of 100,000 records since early this month. Security firms Recorded Future and CrowdStrike have attributed that episode to a Chinese-speaking actor that combined ARTEX with several different AI models.

Separately, Microsoft said the group had used ARTEX since as early as September. Between October 10 and 14, it attempted password-based logins against 327 accounts belonging to eight organizations, generating more than 151,000 failed sign-in attempts without completing authentication.

A spokesperson for the Chinese Embassy in Washington said China "opposes and fights hacking activities in accordance with the law. At the same time, we oppose smears and attacks against China under the excuse of cybersecurity issues." Beijing has consistently denied carrying out or condoning cyber-enabled data theft.

Key details of the campaign, as described by Microsoft:

Note: Timeline and figures are based on Microsoft's account of the campaign.

The disclosure puts a fresh spotlight on the commercial software supply chain that small and mid-sized businesses rely on. Managed service providers aggregate administrative rights across hundreds of client tenants, making them a high-value entry point for espionage-oriented actors seeking scale rather than a single corporate target. ConnectWise's own warning that cleanup at the platform level does not clear intruders already inside customer environments suggests remediation could stretch well beyond the vendor itself.

The role of ARTEX also sharpens a debate over how easily legitimate security tooling can be repurposed. Penetration testing frameworks are standard defensive instruments, but the ability to plug in any AI model and let it run reconnaissance and exfiltration on its own lowers the skill barrier for attackers and compresses the time between initial probing and privileged access - a pattern reflected in the three-hour window Microsoft cited in some intrusions.

For investors, the immediate read-through lands on cybersecurity spending rather than on ConnectWise's privately held balance sheet. Incidents that expose shared service providers tend to accelerate budgets for identity controls, tenant-level monitoring and third-party risk review, particularly among smaller enterprises that outsource IT. Semiconductor, defense and government contractors named among the targeted sectors face elevated scrutiny of their vendor relationships, and cloud identity providers are likely to face louder calls for stronger isolation between customers of the same platform.

Extracted Entities

Attack Types (1)

Companies (1)

Countries (1)

Tools (1)