Chinese Hackers Used an AI Hacking Tool in a Major Supply-Chain Attack, Says Microsoft

A cyber attack revealed this week shows how quickly artificial intelligence is changing hacking. Microsoft says a China state-aligned group it tracks as Storm-0940 used an AI-powered hacking framework to compromise ConnectWise’s Cloud Backup service, then tried to break into more than 2,250 organisations’ Microsoft 365 accounts at machine speed. It is one of the clearest public examples yet of an agentic AI cyberattack, where AI agents automate much of the work human hackers used to do. Here is what happened, how the AI tool was used, and what organisations should do now.

What Microsoft Found

Microsoft Threat Intelligence published its findings on 21 October 2026. It said that as early as September, Storm-0940 used an agentic AI penetration testing framework called ARTEX to automate attacks targeting hundreds of organisations worldwide across many industries, as Microsoft’s security blog explains. Microsoft assessed that the group was behind the recently disclosed compromise of ConnectWise Cloud Backup, a service formerly known as SkyKick.

Microsoft stressed that the activity did not result from a compromise of, or a vulnerability in, its own systems or products.

What Is ARTEX?

ARTEX is an open-source framework built from AI agents powered by large language models. It can work with different AI models and run an automated penetration test from reconnaissance through data theft with little human involvement. Penetration testing is normally used by security teams to find weaknesses before criminals do, but in the wrong hands the same tools can power real attacks.

Researchers have reported that ARTEX configurations pointed at AI models including DeepSeek, GLM and Grok. Public reporting has also linked the framework to attacks on South Korean financial institutions.

How the Attack Worked

The attackers took advantage of trust that organisations had already granted. After compromising ConnectWise’s Cloud Backup application and obtaining its credentials, Storm-0940 used ARTEX to sign in to Microsoft 365 environments as that trusted application. That meant it did not need anyone to log in interactively.

The speed was striking. Microsoft said the group reached more than 2,250 tenants at about 2,900 sign-in attempts per minute, according to The Record. While most of the targeting failed, Microsoft said Storm-0940 still achieved privileged cloud access at many organisations in a short time, in some cases within about three hours.

AI-accelerated activity after entry included reconnaissance, changing application credentials and attempting to raise privileges to keep access.

What ConnectWise Has Done

ConnectWise shut down several services after warnings about the compromise. Reports say existing backup archives remain encrypted, isolated and preserved, with no sign that the backups themselves were compromised.

What Organisations Should Do

If your organisation used ConnectWise Cloud Backup or SkyKick, treat this as urgent. Microsoft and security experts recommend:

• Review trusted applications in Microsoft 365 and remove access that is no longer needed.
• Preserve and review logs for unusual sign-ins and permission changes.
• Rotate credentials and secrets linked to affected applications.
• Check for new or changed permissions that attackers may have added.
• Plan for machine-speed attacks, using automated detection and response rather than relying only on human review.

The same lesson appears across this week’s security news: attackers move fast when credentials and permissions are loose. See our coverage of the PHP2Shell flaws, the Ogi Password1 breach and the US DeepSeek ban. Our cybersecurity tips for businesses guide covers more steps.

Hacking at Machine Speed

The Storm-0940 campaign shows AI agents moving from research demos into real attacks against real organisations. Defenders now face threats that can try thousands of logins a minute. Tighten trusted app permissions, watch your logs closely, and invest in defences that can respond as fast as AI-driven attackers.

Frequently Asked Questions

Who is Storm-0940?

It is a China state-aligned threat actor, as tracked by Microsoft Threat Intelligence.

Was Microsoft hacked?

No. Microsoft says the activity did not stem from a compromise of or vulnerability in its own systems or products.

What is an agentic AI cyberattack?

It is an attack in which AI agents automate tasks such as scanning, testing credentials and expanding access, at far greater speed than humans.

Were ConnectWise backups stolen?

Reports say existing backup archives remain encrypted and preserved, with no sign they were compromised.