
What happened
Microsoft says China state-aligned hackers stole application credentials from ConnectWise's cloud backup service, then used AI agents — specifically the open source ARTEX framework — to attempt sign-ins at almost 3,000 per minute across around 2,250 organizations.
Why it matters
With stolen credentials, the hackers could log in to a customer's environment as an authorized user and access emails and other files, while the sign-in speed gave defenders little time to detect and respond.
WHO IT HITSIT and security teams at the roughly 2,250 organizations that relied on ConnectWise's cloud backup service face credential exposure and must treat their hosted environments as potentially accessed, while ConnectWise's own incident responders work to restore the paused service.
Summaries like this, in your inbox every morning.
The ConnectWise breach stands out less for the intrusion itself than for how the attackers scaled it. According to Microsoft, the hackers did not need to break into each customer individually; they stole the application credentials that ConnectWise uses to link its cloud backup service to corporate customers, then logged in as authorized users. That turned a single breach at an IT management and remote access provider into access to emails and other files across thousands of organizations.
The use of AI agents is what made the campaign fast enough to overwhelm defenses. Microsoft says the hackers attempted almost 3,000 sign-ins per minute across around 2,250 organizations, a pace it says gives defenders little time to detect and respond. The framework named by Microsoft, ARTEX, is an open source AI tool designed to stress-test cyber defenses — the same software Microsoft says was used in a recent compromise of South Korea's financial systems, many of which had weak security and lacked multi-factor authentication.
ConnectWise has paused access to its cloud backup system and, as of Thursday, said it was working on restoring the service. For the customers involved, the immediate question is not only whether the hackers got in, but whether the credentials that made the links work have been fully invalidated — a process that, by Microsoft's account, the attackers were able to outrun.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Morgan Stanley already blocks AI agents entirely, and JPMorgan, Bank of America and Citi are doing the same, A…

Modal CEO Erik Bernhardsson said the AI infrastructure company crossed $1 billion in annualized run rate, rais…

Researchers using AI models including OpenAI's GPT-5.6 Sol discovered two PHP vulnerabilities, collectively na…

Microsoft's Diagnostic Orchestrator, powered by OpenAI's GPT-5, scored about 83 percent diagnostic accuracy on…

Zuckerberg said Muse, released to the public last month, has been downloaded and set up by more than 10 millio…

Modder Sammy Hajhamid, aka "pepsipu," released a software-only, tethered jailbreak called "ns2jb" for the Nint…
