Trellix Advanced Research Center
Between April 1 and September 30, 2026, the most consequential operational technology (OT) development was a cluster of government-confirmed, direct control-system activity against the U.S. water and wastewater sector. Multiple official reporting streams—a CISA advisory on Iranian-affiliated programmable logic controller (PLC) targeting and its July update, Minnesota state reporting on more than 30 community water systems, and a July FBI/EPA warning on internet-facing PLCs—independently documented malicious interaction with PLCs and human-machine interfaces (HMIs).
These are separate reporting streams with different observation windows, affected equipment, and attribution; this report treats them as corroborating a pattern of water-sector PLC exposure, not a single campaign or an additive victim count.
Against that backdrop, Trellix telemetry offers a large, complementary view of the enterprise-IT threats reaching industrial-sector organizations. Across nine industrial-sector categories, it recorded 29.8M malicious-file detections in the period—about 1% of them campaign-associated (matches to threat-intelligence indicators)—concentrated in manufacturing (49.3%), industrial (13.0%), and energy/oil and gas (12.3%). The data reflects which industries generate detections, not which technologies were targeted.
Separately, the vulnerability exploitation we documented in-period centered on enterprise-facing appliances—VPNs, application-delivery controllers, and identity systems that industrial organizations also use for remote access and perimeter defense. These create real risk for industrial remote access and the IT/OT boundary, but the available evidence does not establish that any of these enterprise-appliance compromises progressed into an industrial control environment.
Industrial ransomware remained prominent by one measure only—the volume of leak-site extortion claims. Across five core industrial sectors, Trellix ransomware tracking recorded roughly 1.5K posts, up +50.8% quarter-over-quarter in posting volume.
Finally, the period included at least one documented destructive-OT incident: Incident-response firm Profero reported a dual IT/OT sabotage operation at an Israeli food-production plant in May 2026, including manipulation of refrigeration controllers and destroyed equipment. We present this as an important direct-OT case, attributed explicitly to Profero and uncorroborated by independent Trellix observation.
Internet-exposed PLCs and enterprise remote-access infrastructure remain critical points of exposure for industrial environments. Prioritize removing controllers from direct internet access, rapidly remediating actively exploited edge-appliance vulnerabilities, and strengthening segmentation and coordinated threat monitoring across IT and OT environments.
During the reporting window, Trellix telemetry recorded 29.8M malicious-file detections at organizations across nine industrial-sector categories.
The volume concentrates heavily in manufacturing-aligned industries. Manufacturing alone accounts for 49.3% of the selected detection volume, followed by industrial (13.0%) and energy/oil and gas (12.3%); construction (8.0%), automotive (5.1%), transportation and shipping (3.7%), aerospace and defense (3.7%), utilities (3.6%), and process manufacturing (1.3%) make up the remainder (Figure 1). A high detection share in a sector reflects security-product activity at companies in that industry, not control-system compromise.
Industrial-sector malicious-file detections
Across five core industrial sectors (manufacturing; construction; transportation and warehousing; mining, quarrying, and oil and gas extraction; utilities), Trellix ransomware tracking recorded roughly 1.5K leak-site posts, up +50.8% quarter-over-quarter (Figure 2). Manufacturing accounts for about two-thirds of this population; utilities—the most OT-salient sector—for under 2%. The increase is in posting volume; it may reflect, at least in part, posting cadence and a shifting group landscape (new and newly scaled groups appeared during the window) rather than a demonstrated rise in intrusions. The records carry no post-incident operational detail, so the absence of reported disruption is not evidence that none occurred.
Industrial ransomware leak-site claims, by month
A third, independent view comes from Trellix SecondSight, our threat-hunting service, whose analyst-reviewed notifications to customers offer a small, qualitative complement to the automated telemetry above—not a second volume metric, and not combined with it. Across the period, these notifications spanned ten OT-relevant industrial sectors, echoing the broader pattern: Manufacturing led (~30%), followed by food/agriculture, transportation and shipping, and industrial. Most originated from endpoint (EDR) telemetry (~81%), with the remainder from email and Trellix Helix.
What is useful here is not the count but the character of the threats reaching these organizations. Classified from the notifications, the most common themes were phishing and malicious attachments, remote access trojans (RATs), information stealers, and PowerShell/living-off-the-land activity—with RATs the single most common primary category. Several of these are commonly associated with the initial-access, remote-control, and credential-theft stages of an intrusion; their prevalence at industrial-sector organizations is a reminder of why the IT/OT boundary matters.
The clearest direct-OT signal of the period came from U.S. water and wastewater reporting. Three official streams overlap in time (late July 2026) and sector but are not established by their sources as the same operation, the same actor, or additive victim counts, and we do not present them as such.
The three streams are consistent with a broader pattern of internet-exposed water-sector PLC/HMI exposure in mid-2026.
Incident-response firm Profero reported that, in May 2026, it investigated a dual IT/OT sabotage operation at an Israeli food-production plant. According to Profero, the attackers reprogrammed industrial refrigeration controllers (altering setpoints and safety limits), forced motorized valves to manual and pinned them open, wiped controller configurations, changed engineer credentials to lock out operators, and destroyed three compressors, forcing a multiday recovery. A Go-based toolkit ("GRAT") was used, and an IT-side disk-wiper component was staged but not executed.
Profero attributes the operation with high confidence to Cyber Isnaad Front, an Iranian state-directed persona associated with Aria Sepehr Ayandehsazan (ASA). This account is Profero's alone—not independently corroborated by Trellix telemetry or a second source—and the initial-access method and IT-to-OT movement path are not documented; we do not infer them. It is an important example of hands-on OT sabotage outside the energy/grid context, and one documented destructive-OT incident of the period, not necessarily the only one.
In the five-sector industrial leak-site population, the two most active groups—"The Gentlemen" and Qilin—each accounted for about 16% of posts, well ahead of LockBit5 (~6%), DragonForce (~4%), and Clop (~3%). Claim volume reflects posting behavior, not severity: "The Gentlemen," the top poster, is characterized by high-volume, templated victim listings.
One methodologically notable case is the Indonesian state energy company Pertamina, claimed by two different groups three months apart (The Gentlemen, July 31; RansomHouse, September 17)—the dataset cannot distinguish one double-claimed intrusion from two separate events. Separately, government advisories documented criminal ransomware with industrial-sector victims but no OT-specific tooling: the Gunra ransomware-as-a-service operation (CISA/FBI AA26-222A), exploiting FortiOS vulnerabilities for corporate-IT access; and Cl0p/TA505, exploiting a PTC Windchill/FlexPLM zero-day (CVE-2026-12569) to steal engineering/product-lifecycle data—adjacent to industrial design workflows but not control networks.
Threat activity affecting industrial organizations during April–September 2026 highlighted the intersection of conventional enterprise cyber threats and attacks targeting industrial control systems. Trellix threat intelligence and Trellix SecondSight notifications identified phishing, malicious scripting, remote-access malware, and credential-related threats, while government advisories and incident-response investigations documented direct manipulation of PLCs, industrial configurations, and physical processes.
The following MITRE ATT&CK mapping illustrates these complementary threat patterns, distinguishing enterprise-facing techniques associated with Trellix intelligence from control-system behaviors documented in reported OT incidents.
Enterprise—Phishing
Most common Trellix telemetry observed theme: phishing and malicious attachments
Enterprise—User Execution: Malicious Copy and Paste (ClickFix)
ClickFix/fake-CAPTCHA copy-paste lures seen in Trellix telemetry
Enterprise—PowerShell
PowerShell/living-off-the-land was a recurring theme in Trellix telemetry
Enterprise—Exploit Public-facing Application
Trellix threat intelligence associates exploitation of internet-facing applications and appliances with industrial-sector activity; corroborated by confirmed enterprise-edge exploitation this period
ICS—Manipulation of View
Iranian-affiliated actors manipulated data shown on HMI and SCADA displays.
ICS—Modify Parameter/Modify Alarm Settings
Attackers altered refrigeration-controller setpoints and safety limits (Israeli food-production incident).
ICS—Change Credential
Engineer credentials changed to lock operators out. Unauthorized changes to device passwords on internet-facing PLCs.
ICS—Manipulation of Control
Motorized valves forced to manual and pinned open.
Unauthorized PLC changes with operational effects including pressure loss and flooding.
ICS—Loss of View / Loss of Control
Operators locked out of controllers; multiday, hands-on manual recovery required.
Enterprise mappings are Trellix threat-intelligence associations, not verified technique execution; ICS mappings are behaviors documented in external incident reporting.
The period's vulnerability picture divides cleanly into three bands: native OT/ICS disclosures, enterprise-edge technologies under active exploitation, and the architectural IT/OT-boundary risk those edge technologies create. We keep distinct disclosed, confirmed-exploited; verified operational consequence; and potential operational consequence.
Two OT vendor/CISA disclosures in September 2026 were tracked, neither with observed in-the-wild exploitation:
AVEVA Pipeline Integrity Monitor—CVE-2026-81821 (PIMBoards component), CVSS v4.0 8.3: Disclosed September 8, 2026 (AVEVA-2026-006), this documented weakness lets an attacker with read access to PIMBoards project files potentially decrypt sensitive information.
Schneider Electric SCADAPack x70 RTU—CVE-2026-81861: Insufficiently protected credentials (CWE-522) that could expose authentication information and allow unauthorized access to RTU functionality; CVSS v3.1 6.5. Schneider published SEVD-2026-251-03 on September 8, 2026; CISA issued the corresponding advisory ICSA-26-258-04 on September 15, 2026.
Confirmed exploitation in-period involved enterprise-facing appliances that industrial organizations also deploy for remote access and perimeter defense:
Enterprise VPNs, application-delivery controllers, and identity systems frequently support remote access to industrial environments, making them important points of exposure at the IT/OT boundary. The confirmed exploitations described in this section primarily involved enterprise-facing systems, but these technologies can also provide pathways toward operational networks. Organizations should prioritize securing remote access, enforcing segmentation, and monitoring suspicious activity across IT and OT environments.
Water-sector PLC exposure is a recurring, learnable pattern. The 2026 water-sector reporting echoes the 2023–2024 campaign in which Iran-linked actors exploited internet-exposed Unitronics PLCs across U.S. water utilities using default or weak access—defacing HMIs and altering controller logic. The mechanism then and now is the same: controllers reachable from the public internet via standard engineering access, compromised without a software vulnerability.
Treat industrial ransomware as a business-continuity risk. The period's leak-site data shows a high volume of extortion claims but carries no post-incident operational detail, so it cannot tell us where the harm landed—on business systems, on operations, or not at all. Rather than infer a typical location or impact from records that lack it, the durable planning lesson is that a ransomware event at an industrial operator can threaten business continuity, and interdependencies between IT and operations can force a precautionary operational pause even when control systems are never touched. A 2025 retrospective example—the Blue Locker ransomware incident at Pakistan Petroleum, an IT-only server-encryption event that halted financial operations for roughly two days—illustrates that IT/operations coupling; it predates this reporting window and is included only as a dated reference.
Capability is outpacing demonstrated impact. Several 2026-reported items were capability or attempted activity, not incidents: The ZionSiphon water-targeting malware sample was non-functional (a bug in its own targeting logic); a "Lotus Wiper" sample aimed at the energy-sector naming convention was analyzed with no confirmed deployment; and an AI-assisted reconnaissance effort against a Mexican water utility (late 2025–early 2026) reached but failed to breach an ICS-monitoring interface.
Each recommendation below follows from a specific finding in this report.
The April–September 2026 period draws a sharp line between what was confirmed and where it was confirmed. The strongest OT evidence—government-attributed and incident-response-documented activity against water-sector PLCs and against an Israeli food-production plant—came from external primary sources, and even there the reporting is careful: Operational consequences are real but scoped ("in some cases;" "not every affected community"), attribution is mixed, and distinct reporting streams should not be collapsed into one campaign.
Trellix threat intelligence provides a complementary perspective on enterprise-facing threats affecting industrial organizations, including malicious-file detections, ransomware extortion activity, and vulnerabilities in technologies supporting remote access to industrial environments.
Together, these findings reinforce two immediate priorities: eliminating unnecessary internet exposure from industrial controllers and securing the enterprise infrastructure that provides access to OT environments. Strong segmentation, integrated IT/OT monitoring, and tested recovery procedures remain essential to reducing the risk of operational disruption.
Trellix telemetry and ransomware tracking and Trellix SecondSight findings are proprietary; exact figures, methodology, and source validation are held internally.
This document and the information contained herein describe computer security research for educational purposes only and the convenience of Trellix customers. Trellix conducts research in accordance with its Vulnerability Reasonable Disclosure Policy I Trellix. Any attempt to recreate part or all of the activities described is solely at the user’s risk, and neither Trellix nor its affiliates will bear any responsibility or liability.
Trellix is a trademark or registered trademark of Musarubra US LLC or its affiliates in the US and other countries. Other names and brands may be claimed as the property of others.