Chinese hackers spied on thousands of companies after breaching IT giant ConnectWise
Oct 22, 2026, 9:47 AM · TechCrunch

One stolen set of backup credentials let state-aligned hackers walk into about 2,250 organizations, and AI agents are why it happened at nearly 3,000 sign-ins a minute.
Why it matters
According to Microsoft, China state-aligned hackers broke into ConnectWise's cloud backup service and stole application credentials that link ConnectWise to its corporate customers. With those, they could log in to customers' Microsoft-hosted environments as authorized users and reach emails and files.
Microsoft says the attackers used AI agents to hit around 2,250 organizations at a rate of almost 3,000 attempted sign-ins per minute, leaving defenders little time to respond. ConnectWise paused access to the backup system and, as of Thursday, said it was working to restore service.
From the desk
This is the supply-chain nightmare security teams have warned about, with a new accelerant. Breaching one trusted IT provider has always been the efficient way into thousands of companies. What changes here is the tempo. Microsoft attributes the speed to ARTEX, an open-source AI framework built to stress-test defenses, which reportedly was also used in a recent compromise of South Korean financial systems, many of them lacking multi-factor authentication.
We want to be careful not to blame the tool for the intent. Red-team frameworks exist because defenders need them, and AI-assisted testing is genuinely useful. But this is the pattern we keep seeing this month: the same automation that helps defenders find weaknesses collapses the time attackers need to exploit them. When attempts arrive at thousands per minute, a human-paced incident response playbook is already behind.
The lesson for most organizations isn't exotic. Long-lived application credentials held by third parties are a standing risk, and they deserve the same scrutiny as employee passwords: scoped narrowly, rotated, and watched for unusual sign-in volume. Customers who rely on ConnectWise backups also now face the uncomfortable fact that the service meant to protect their data was the door.
Where this leads if it scales is clear enough. Expect more attackers to target service providers and integration tokens, and expect AI agents to turn every stolen key into a mass campaign within minutes. Defenders will need automated detection and automatic credential revocation that move at the same speed. I'm watching how quickly Microsoft and ConnectWise can tell affected customers exactly what was accessed.
Context
This lands in an already rough week for security, following a widely reported spate of incidents including newly disclosed PHP vulnerabilities and a cyberattack that led developer platform Railway to urge all users to revoke secret keys. ConnectWise is a major provider of IT management and remote access tools used by managed service providers and businesses.
Who feels it
- ConnectWise customers
- Review Microsoft tenant sign-in logs and revoke or rotate any credentials tied to the backup integration while the investigation continues.
- Managed service providers
- Trust in third-party backup and remote access links is under pressure; expect customers to ask hard questions about credential scoping.
- Security teams
- Attack speed of nearly 3,000 attempts per minute argues for automated detection and response on sign-in anomalies.
- AI security tool builders
- Open-source offensive frameworks face growing scrutiny as they show up in real state-aligned campaigns.
What to watch
- ConnectWise's disclosure of what data was accessed and when backup service is fully restored
- Further Microsoft guidance or indicators tied to ARTEX-driven campaigns
- Whether regulators or customers push for limits on long-lived third-party application credentials
- Follow-on breaches at downstream organizations using the stolen access
Companies: Microsoft