AivexaNewsSearch
AI news for builders and product teamsChecked every hour

Chinese hackers spied on thousands of companies after breaching IT giant ConnectWise

Collected Oct 22, 2026

Microsoft's security researchers attributed a campaign against ConnectWise, an IT management and remote access provider, to hackers aligned with the Chinese state. According to the findings reported by TechCrunch's Zack Whittaker, the intruders got into ConnectWise's cloud backup service and took application credentials — the secrets used to establish data links between the service and the corporate customers who rely on it to back up company data. The compromise gave the attackers a way into customer accounts at scale, which is why the affected population runs into the thousands of organizations.

The mechanics matter for anyone running a managed service provider stack. Because the stolen credentials were tied to the connection between ConnectWise and customer environments, using them let the hackers authenticate to customers' Microsoft-hosted environments as though they were an authorized user. From there they could reach email and other files. Microsoft characterized the activity as spying, so the concern is data access and persistence rather than disruption — though any stolen credential of this kind can be reused long after the initial entry.

What stands out is velocity. Microsoft said the attackers leaned on AI agents to push into roughly 2,250 organizations, with attempted sign-ins arriving at close to 3,000 per minute. Microsoft's own framing is that this rate leaves defenders with very little time to detect and respond. In practice, that arithmetic is the whole problem: a defender watching authentication logs has minutes, not days, before an automated campaign has cycled through an entire target list.

The AI tooling named in the report is ARTEX, described as an open-source AI framework that pairs with AI models to stress-test cyber defenses. Microsoft linked the same software to a recent compromise of South Korea's financial systems, where many institutions reportedly had weak security and no multi-factor authentication. That second detail is worth sitting with: the South Korea case suggests the framework works best against soft targets, and multi-factor authentication remains the cheapest, bluntest control against stolen-password reuse.

The defensive reading here is not that AI has invented a new attack class. Credential theft and mass authentication attempts are old; automation simply compresses the timeline. What changes is the ratio of attacker throughput to defender review capacity. A SOC that triages login anomalies in human time will lose ground against a scripted campaign operating at thousands of attempts per minute. Rate limiting, anomaly detection tuned for authentication bursts, and short-lived scoped credentials for service-to-service links all become more important than ever, and a stolen link credential should be treated as a standing exposure rather than a one-time incident. This analysis is my own reading of the reported facts, not a statement from Microsoft or ConnectWise.

ConnectWise has said it paused access to its cloud backup system and is investigating. As of Thursday the company said it was working to restore the service. No timeline for full restoration was given in the report, and the exact scope of what the intruders read at each of the roughly 2,250 organizations is not public.

Why it matters: Managed service providers are a chokepoint — a single compromise at an IT vendor multiplies across every customer that trusts its integration. Organizations using ConnectWise's cloud backup should assume their link credentials were in scope, rotate them, and check Microsoft-hosted environments for sign-ins that don't match normal patterns. It is a fair inference, not a reported finding, that AI-assisted attack tooling will push vendors and defenders alike toward faster automated detection and tighter credential lifetimes.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Microsoft points the blame for ConnectWise's breach at a China-aligned hacker, which relied on AI agents to rapidly hack into thousands of its customers.