OTSecurity

Claroty finds 65,000 consumer devices connected to critical infrastructure networks, exposing OT segmentation gaps


Claroty found more than 65,000 non-essential consumer devices connected to networks supporting critical infrastructure, including OT (operational technology), healthcare and commercial environments, raising concerns about network segmentation and cybersecurity governance. The findings are based on an analysis of more than 45 million devices across 30,000 critical infrastructure sites. The inventory included 16,745 gaming consoles, most located in hospitals, 44,059 smart home and entertainment devices, and 4,915 connected vehicles operating on OT networks.

In its research titled ‘Game Over for Consumer Tech Risk in Critical Infrastructure,’ Claroty identified security weaknesses associated with unmanaged consumer technology, including unpatched firmware, outdated operating systems, known remote code execution vulnerabilities and connections to external, vendor-controlled cloud services. Claroty found that 95% of gaming consoles in healthcare environments were connected to production networks rather than isolated guest or IoT networks. 

Almost none of the connected vehicles analyzed were properly isolated, while researchers identified six critical CVEs affecting Tesla Model 3 vehicles and 121 security issues associated with Ford F-150 trucks. The findings point to deficiencies in network segmentation, asset visibility and access controls that could provide threat actors with pathways into mission-critical environments, particularly as AI lowers barriers to reconnaissance and vulnerability scanning. 

Additional findings showed that 95% of gaming consoles in healthcare environments were connected to production networks rather than isolated guest or IoT networks or VLANs. By contrast, personal electronics such as smartwatches and smart speakers demonstrated better network hygiene, with 71% appropriately connected to guest networks. Almost none of the connected vehicles analyzed were properly isolated on guest networks. Researchers also identified six critical CVEs affecting Tesla Model 3 vehicles and 121 security issues associated with Ford F-150 trucks.

“Finding consumer technology like Apple Watches and Nintendo Switches residing on the same networks as OT and medical equipment points directly to ongoing deficiencies in network governance and segmentation within operational settings,” Sean Tufts, field CTO at Claroty, said in a media statement. “Addressing this issue does not mean stripping smart TVs from conference rooms or banning XBOX consoles from hospitals; it requires closing off unnecessary pathways for malicious actors from causing intolerable disruptions and negatively impacting businesses across industries.”

Despite millions of dollars invested in technology to protect critical infrastructure and build operational resilience, the explosion of connected devices has led to failures in virtual network segmentation, continuous threat detection, remote access, and overall governance. At a time when AI is lowering the barrier to entry even further for threat actors to not only carry out reconnaissance but also scan internet-facing assets for known vulnerabilities, the number of devices left to provide a foothold for attackers was staggering. 

Organizations must adopt rigorous access controls and network segmentation, along with comprehensive visibility into connected assets, to safeguard the resilience of key operational systems. Enforcing continuous asset discovery and strict segmentation between guest, IoT, and core production environments eliminates unmanaged entry points and prevents threat actors from moving laterally to compromise mission-critical infrastructure.

Claroty’s research uncovered a consequential number of consumer technology devices such as smart home and entertainment systems, gaming consoles, and cars connected to sensitive operations networks. These devices may appear harmless, but when they share network infrastructure with ICS (industrial control systems), building management systems, and medical devices, they can become an entry point for attackers seeking to disrupt production, compromise critical processes, or threaten operational resilience. 

Consumer devices are rarely designed to meet the security, reliability, and lifecycle requirements of operational environments. Software and firmware vulnerabilities are present on these devices, many of which are not secure by design or default. Many have insecure communications or must reach out to the public internet to a cloud service that is outside a factory’s or hospital’s control. 

Most organizations don’t have visibility into these rogue third parties’ security posture and risks they may introduce to production and medical networks. So clearly, the fundamental problem is that these devices introduce unmanaged assets, unexpected communications, and potential vulnerabilities into an environment where a compromise can have physical consequences. 

Of the more than 45 million devices analyzed, over 450,000, or approximately 3.5%, were classified as critical-risk devices because of known vulnerabilities, outdated firmware or a lack of security updates. The dataset covered PCs and tablets, automation equipment such as programmable logic controllers (PLCs), controllers and human-machine interfaces (HMIs), industrial printers, IP cameras and surveillance systems, as well as robotics and manufacturing systems.

Apple had the highest critical-risk devices, with 155,972, followed by Rockwell Automation with 132,868 and Axis Communications with 40,706. Other vendors included Schneider Electric with 28,568 devices, Zebra Technologies with 19,068, Siemens with 13,420, Hanwha Techwin with 12,775, Fanuc Robotics with 7,332, Honeywell with 4,335 and GE with 4,326.

The research also addressed that device risk is not only determined by a CVE or CVSS score; multiple factors make up scoring criteria. It considers presence of confirmed or potentially relevant common vulnerabilities and exposures (CVEs), particularly those with high Common Vulnerability Scoring System (CVSS) scores, known exploits listed in CISA’s Known Exploited Vulnerabilities catalog, or high Exploit Prediction Scoring System (EPSS) scores. It also examined whether devices communicated with the internet, particularly through inbound or server connections; risks associated with outdated, unsupported, or end-of-life operating systems; and weaknesses in security controls, including missing endpoint protection, unmanaged devices, or missing security updates. The assessment also considered each device’s operational importance and potential impact of its compromise.

Highlighting recent attacks on European supermarkets, for example, Claorty identified that attackers are accessing networks via unsecured internet-facing assets. “In the supermarket attacks, attackers were able to manipulate temperature controls and defrost food, causing millions in lost stock and revenue. Once a device is compromised, its position on the network becomes important. If it can communicate freely with other systems, an attacker may attempt to discover reachable assets, abuse improperly configured access permissions, steal credentials, or exploit vulnerabilities in adjacent computers and servers. A compromised device can also provide a foothold for further reconnaissance and lateral movement.” 

The report observed that the greatest danger emerges when connectivity between IT and OT networks is insufficiently controlled. Determined attackers looking to manipulate or disrupt operations can exploit weak or non-existent segmentation, excessive permissions, shared credentials, or unmonitored connections to move deeper onto the network. An innocuous connection from a connected automobile on the same network as OT, IoT, and BMS assets can provide a pathway to access those sensitive devices. If the pathway is unknown, the risk is substantial. 

Claorty noted that the problem for OT, commercial, and healthcare networks is that they may operate outside IT’s control, with individual sites maintaining autonomy over their operations. This decentralized approach has created an architecture characterized by perimeter-focused security, often relying on a single large boundary firewall with minimal internal controls; limited network segmentation, with few guest networks or isolated VLANs for non-critical devices; minimal access controls, allowing devices to connect to corporate OT networks with few barriers; and mixed-use infrastructure, where consumer devices share network space with human-machine interfaces (HMIs), programmable logic controllers (PLCs), magnetic resonance imaging (MRI) systems, and cooling equipment.

The report further detailed, “We not only found tens of thousands of consumer devices on OT networks, but also exploitable CVEs on the consoles, entertainment devices, and cars that could provide determined attackers with a foothold on the network.”

Claroty noted that segmentation has become the default answer for networks where downtime is simply not an option, whether that means a production line, a hospital ward or a commercial supply chain. Yet the gap it is meant to close keeps widening, and Claroty’s analysis points to three stubborn causes. The first is decentralized management. When individual sites can write, bend or ignore security rules on their own, a uniform posture dissolves into patchwork: patches land late, tool sets multiply, and the central security operations center loses sight of the very locations attackers find easiest to breach. 

The second is legacy architecture. Networks built decades ago around a perimeter mindset no longer match how industrial environments actually behave. Digital transformation has pushed data sideways, across layers and out to cloud and partner systems, which cuts against the vertical logic of the Purdue Model and lets intruders move laterally with little resistance. Underneath it all sits a cultural divide, since OT prizes availability and safety while IT leans on confidentiality and tolerates planned outages, and that friction makes it hard to apply standard frameworks across converged environments.

The third cause, slow modernization, is where the practical pain lies. Many of these assets are out of vendor support yet remain woven into production or patient care, so ripping them out is costly and risky. Every change has to account for hidden dependencies, custom configurations and critical traffic that cannot be interrupted, because a clumsy fix can do as much damage as an attack. 

Even routine remediation, from firmware updates to configuration fixes, demands scheduled downtime, regression testing and close monitoring once changes go live. Add the expense of new hardware, software and scarce specialist expertise, and the result is a long window of exposure to known exploited vulnerabilities, weak configurations and insecure remote access. In short, the problem is less a shortage of security tools than a collision between operational reality and the pace at which defenses can safely change.

Claroty argued that the security of hospitals, power plants, manufacturing facilities and water treatment systems depends directly on how rigorously organizations enforce network segmentation. Every Nintendo console, Tesla vehicle or Apple Watch connected to a production network expands the attack surface, creates a potential pivot point for lateral movement, introduces a blind spot in security monitoring and indicates a failure to enforce segmentation policies.

To strengthen critical infrastructure security, organizations should implement strict network segmentation by separating guest, IoT and production networks through enforced security policies. They should deploy continuous asset discovery to maintain real-time visibility into all connected devices and enforce secure access requirements to prevent unauthorized devices from accessing production networks. Claroty also recommended accelerating modernization efforts by replacing legacy systems and redesigning network architectures, while establishing governance frameworks that centralize security policies without undermining operational autonomy.



Source link