The lights can stay on while someone learns how to turn them off.
That was the fear behind a warning US investigators issued in February 2024. They had found signs that intruders had maintained access to some critical-infrastructure networks for at least five years. The affected sectors included energy and water, services most people think about only when they stop working. Joint cybersecurity advisory, AA24-038A
The campaign was called Volt Typhoon. US agencies attributed it to Chinese state-sponsored hackers and assessed that they were preparing access for disruption during a future crisis.
Think about what that would mean during a confrontation: the people responsible for keeping essential services running could also be trying to work out who else had access to their systems. Any resulting disruption would arrive when those services were already badly needed.
The advisory documented intrusions, not that feared catastrophe. Its warning concerned what the access might allow someone to do later.
Five years is a long time to keep a way into someone else's network. What were they learning while they were there?
They could use the tools already inside
Microsoft publicly described the campaign in May 2023. Its researchers had observed activity targeting organizations in Guam and elsewhere in the United States. They assessed, with moderate confidence, that the campaign was developing capabilities that could interrupt critical communications between the US and Asia during a future crisis. Microsoft's original investigation
The methods helped explain how an intrusion could last. Microsoft described hackers using stolen credentials and tools already present on the computers they entered. Security researchers call this “living off the land.”
A familiar administrative program can serve a legitimate technician or an intruder. A stolen account can pass a password check. Defenders then have to recognize that the person using it does not belong there, even though the software and login appear legitimate.
Microsoft also reported traffic routed through compromised home and small-office networking equipment. That helped conceal the operators among ordinary internet activity. Someone's router could be carrying traffic toward a critical target while its owner continued using it without knowing anything was wrong.
The routers the FBI took back
In December 2023, a court-authorized US operation targeted part of that concealment system. The Justice Department announced the action the following month: hundreds of US-based routers had been hijacked into what investigators called the KV Botnet. Justice Department account of the operation
Most were Cisco and Netgear devices that had reached the end of manufacturer support. Security patches were no longer arriving. Investigators said Volt Typhoon used the infected routers to disguise the Chinese origin of further hacking activity.
The operation deleted malware and severed connections to the botnet. It removed equipment the hackers depended on, but the announcement did not establish that every intrusion at every downstream target had been eliminated.
That leaves a practical problem after a takedown. Removing a route an intruder used to get somewhere is different from establishing that they have lost every account or foothold they acquired after arriving.
The route toward physical control
The February advisory also confirmed compromises in communications and transportation, including in Guam. US agencies assessed with high confidence that the targets and behavior pointed toward preparing movement into operational technology: systems involved in physical processes.
In one confirmed intrusion, the actors reached a control system and were positioned to move toward another. The report does not establish an ability to trigger a nationwide blackout. It does show why investigators were concerned about access extending toward equipment with a physical job to do. Technical findings in the joint advisory
A subsequent leadership briefing made the agencies' concern explicit: access could enable disruption or destruction of critical services during heightened geopolitical tensions or military conflict. It urged infrastructure leaders to treat the threat as a business risk requiring action. Joint fact sheet for infrastructure leaders
The consequences would depend on the equipment within reach and the permissions the intruders held. Operators could detect the activity or interrupt an attack. Those barriers matter, because access alone cannot tell us how much damage someone could cause.
The risk is that defenders discover the limits of an intrusion only after someone tries to exploit it. A network that appears to be working gives the public little reason to suspect that essential services may be exposed.
China disputes the attribution. In February 2024, its foreign ministry rejected the allied advisory as a fabricated accusation. That denial stands against the technical findings and assessments published by Microsoft and the government agencies. China's official response
Why the warning survived the takedown
In April 2026, the UK's National Cyber Security Centre and international partners returned to the broader problem of covert networks built from compromised devices. Their advisory cited Volt Typhoon's use of such networks to prepare offensive capabilities against critical infrastructure. It described networks that change as devices are removed and others are added. April 2026 joint advisory
Disabling a botnet disrupts a particular operation. The technique remains available wherever attackers can compromise replacement devices. The April warning does not tell us which earlier victims remain compromised today.
For an infrastructure operator, that uncertainty creates an ugly question: has the intruder been removed, or has the investigation only found part of the route they used?
US agencies believe Volt Typhoon was preparing for a crisis in which disrupting civilian infrastructure could become useful. The records examined here supply no attack date or authenticated order to strike. They cannot settle how much access survived or what damage it could enable.
The warning asks us to consider a threat that can exist while the taps still run and the lights still come on. An intruder who wants access for a future emergency has a reason to avoid an obvious failure today. Normal service, by itself, cannot tell the people depending on it who else is inside.







