Minnesota Water Attacks and the Predictable Reality

Minnesota Water Attacks and the Predictable Reality

Minnesota Water Attacks and the Predictable Reality 1200 630 Andrew Heller

In late July 2026, the critical infrastructure community was jolted by a coordinated cyber intrusion affecting more than 30 municipal water systems across Minnesota (Reuters). The attack targeted the operational technology (OT) layer, specifically programmable logic controllers (PLCs) and associated human-machine interfaces (HMIs), forcing utilities into degraded, manual operations (ABC News).

Preliminary intelligence points to Iran-linked actors, though formal U.S. government attribution remains pending (AP News). What matters more for defenders is the architecture that made this possible.

At Lares, we often say attackers don't hack in. They log in.
The Minnesota campaign is a textbook example of adversaries capitalizing on easily accessible, low-hanging fruit rather than burning zero-day exploits.

Here is a technical breakdown of the incident, the underlying architectural failures, and how adversarial engineering teams can proactively test their environments to prevent a similar fate.

Incident Timeline

Date

Event

Source Confidence

Jul 26 to 27, 2026

Coordinated intrusion affects 30+ MN water systems; MNIT confirms malicious activity

Confirmed (Reuters)

Jul 28, 2026

MNIT publicly discloses attack; scope grows to ~36 systems

Confirmed (Wikipedia; AP wire via Yahoo)

Jul 29 to 30, 2026

CISA issues warning; officials report Iran as leading suspect

Strongly Supported (NYT; CBS News)

Jul 30, 2026

FBI confirms awareness, does not attribute

Confirmed (AP News)

Jul 31 to Aug 1, 2026

Leaked WaterISAC/MN Fusion Center memo ties activity to CISA AA26-097A Iran campaign

Strongly Supported, unofficial source (Wired)

Aug 3 to 4, 2026

Broader reporting confirms multi-state impact; no contamination confirmed anywhere

Confirmed (The Guardian)

A note on attribution discipline: as of early August, no agency had formally attributed the Minnesota event. A former senior FBI official acknowledged the possibility of a false-flag operation mimicking Iranian tradecraft, though they characterized it as unlikely given the totality of evidence (Yahoo/AP wire). It's also worth resisting the urge to collapse the 2023 CyberAv3ngers campaign, the April 2026 CISA advisory, and the July 2026 Minnesota event into a single operation. They share tradecraft patterns. They are not confirmed to be the same actors, tools, or campaign.

The Incident: Speed, Scope, and the Power of Manual Fallback

Between July 26 and 27, 2026, approximately 36 community water systems in Minnesota experienced unauthorized interactions with their monitoring and control equipment. The targeted environments included cities like Plymouth, South St. Paul, Maple Plain, and Braham (AP wire via Yahoo).

Unlike financially motivated ransomware events such as the 2021 Colonial Pipeline attack, the threat actors did not demand payment or exfiltrate data (Wikipedia). Instead, they focused on opportunistic disruption. In Braham, malware disabled automated plant controls, leaving the water tower unable to fill for about 90 minutes (AP wire via Yahoo).

Minnesota officials stated there were no active requests for residents to modify water use in the initial disclosure. CISA's broader advisory on this campaign does reference boil-water notices and sustained manual operations occurring elsewhere in the wider multi-state activity, an important distinction between what was confirmed locally and what was reported regionally (AP News).

Human operators were the ultimate fail-safe. They isolated the systems, loaded known-good backups, and restored services manually, preventing any water contamination or safety hazards.

Technical Anatomy of the Intrusion

While formal U.S. government attribution is pending, the technical hallmarks of the attack align with previously documented Iran-linked tradecraft, specifically the April 2026 CISA advisory (AA26-097A) and the 2023 CyberAv3ngers campaign against the Municipal Water Authority of Aliquippa (CISA AA23-335A).

From an adversarial perspective, the attack paths relied on recurring preconditions:

Internet-Exposed PLCs and HMIs

Just as the Aliquippa attack exploited an internet-facing Unitronics PLC on TCP port 20256 with the default password "1111" (RISI database), the Minnesota attacks likely leveraged similarly exposed remote-access vectors. CISA advisories linked to this activity flag exploitation on ports 44818, 2222, 102, and 502 (CISA AA26-097A).

The Cellular / IoT Gateway Vector

In Plymouth, attackers reportedly compromised equipment reachable via cellular communications at two water towers and multiple lift stations (AP wire via Yahoo). This represents a bypass of the traditional IT perimeter, directly exposing field devices to the internet if cellular gateways lack strict authentication or anomaly monitoring.

Vendor-Concentration Risk

Analysts flagged Rockwell Logix controllers vulnerable to CVE-2021-22681, which allows unauthenticated access and lacks a vendor patch, as a plausible vector for part of the campaign (AP wire via Yahoo). When a single vendor's default configuration or unpatched CVE is widespread, it creates a massive, simultaneous attack surface across independent utilities.

Comparative Analysis: Water-Utility Incidents

Incident

Year

Actor

Initial Access

Impact

Confidence

Oldsmar, FL

2021

Unattributed

Dormant TeamViewer account, shared password, no firewall, Windows 7 (CNN; Ars Technica)

Attempted sodium hydroxide overdose; caught and reversed by operator in minutes; no health impact (CNN)

Confirmed technical facts; motive/actor unknown

Aliquippa, PA

2023

CyberAv3ngers (IRGC affiliated)

Internet-exposed Unitronics PLC, default password "1111," default port 20256 (RISI database)

HMI defaced; booster station pressure control forced to manual; no water quality impact (CNN)

Confirmed, jointly attributed by CISA/FBI/NSA/EPA

Minnesota

2026

Suspected Iran-linked, unconfirmed

Unknown; plausible overlap with default-credential/internet-exposed PLC pattern (CISA AA26-097A)

30 to 36 utilities affected; manual operations at several; no confirmed contamination (AP wire via Yahoo)

Strongly supported for activity; Plausible for attribution

Two adjacent critical infrastructure cases reinforce the same lesson. Colonial Pipeline (2021) was traced to a single MFA-less legacy VPN credential, forcing a precautionary shutdown of the largest East Coast fuel pipeline without confirmed direct OT manipulation (CNBC). Volt Typhoon, a PRC state-sponsored actor detailed in a February 2024 joint CISA/TSA advisory, uses no malware at all, living entirely off legitimate administrative tools already present in energy, water, and communications environments to pre-position rather than disrupt (TSA/CISA joint advisory).

The Core Failure: Bypassing the DMZ

The generalized trust-boundary model for an OT environment mandates a mediating DMZ/jump-host layer that separates the IT network from the control network (SCADA/historians) and the field boundaries (PLCs/RTUs).

The single most consequential architectural failure in both the Minnesota and Aliquippa incidents was that the field and control boundaries were reachable directly from the internet, completely bypassing the intended DMZ layer.

Whether through legacy remote-access tools like TeamViewer (as seen in Oldsmar in 2021), MFA-less VPNs, or direct cellular telemetry links, adversaries did not have to move laterally through a segmented corporate network. They went straight for the target.

An Adversary's Perspective: Why Burn Expensive Exploits?

Our take on this event highlights a stark reality: an attacker seeking disruptive access will prioritize low-cost, high-visibility targets over sophisticated tradecraft. This is the same logic we use when scoping engagements for clients. Why would an adversary spend resources developing a novel exploit when a default password on an internet-facing PLC gets the job done just as well?

MITRE ATT&CK Mapping

Technique

ID

Evidence Basis

Exploit Public-Facing Application / Internet-Accessible Device

T1190 and analogous ICS techniques

Confirmed pattern: Aliquippa, CISA AA26-097A (CISA)

Brute Force: Password Guessing / Default Credentials

T1110

Confirmed: Unitronics default password (CISA AA23-335A)

Valid Accounts

T1078

Plausible for Minnesota; Confirmed for Oldsmar shared TeamViewer credential (Ars Technica)

Modify Controller Tasking / Unauthorized Command Message (ICS analog)

Analogous to ICS T0855 / T0836

Strongly Supported: malicious project file interaction, HMI manipulation (CISA AA26-097A)

Living Off the Land / Native tool abuse

Multiple LOLBin techniques

Confirmed for Volt Typhoon; Unknown for Minnesota (TSA/CISA advisory)

Data Encrypted for Impact / Ransomware

T1486

Confirmed for Colonial Pipeline; explicitly not observed in Minnesota; no ransom demand (Wikipedia)

When we run these mappings against our own client environments, we rarely need to reach past the first three rows of this table. Adversaries do not need to emulate the stealthy living-off-the-land pre-positioning of a group like Volt Typhoon if their goal is immediate, attention-grabbing disruption. Scanning the internet for exposed PLCs and trying default credentials reliably generates operational disruption at near-zero cost to the attacker. That is exactly the mindset we bring into every OT-focused engagement: think like the cheapest path in, not the most impressive one.

Risk Matrix and Prioritized Findings

Finding

Likelihood

Consequence

Detectability

Confidence

Internet-exposed PLC/HMI with weak or default credentials

High

High

Low

Confirmed pattern

MFA-less remote-access accounts (VPN, TeamViewer-class tools)

High

High

Low

Confirmed pattern

Lack of IT/OT segmentation gateway

Medium to High

High

Low

Strongly supported

Vendor-concentration / shared-CVE exposure across many utilities

Medium

Medium to High

Low

Plausible to Strongly supported

Absent OT-aware monitoring and telemetry

High

Medium

N/A

Confirmed pattern

Manual operations and backup readiness gap

Variable by utility

High if absent

N/A

Strongly supported where present

Defensive and Adversarial Engineering Takeaways

The recurrence of these exact exposure patterns across 2021, 2023, and 2026 indicates that organizational and architectural hygiene is lagging, not technology. These findings map directly to controls already outlined in the NIST Cybersecurity Framework 2.0, CISA's Cross-Sector Cybersecurity Performance Goals, and IEC 62443 segmentation principles (CISA CPGs). The gap is not guidance. It is implementation.

1. Immediate Architectural Fixes

  • Conduct an immediate inventory of all internet-facing OT devices and disconnect any PLC or HMI directly reachable from the public internet.
  • Rotate all default and shared credentials, and enforce MFA on all remote-access and OT-adjacent accounts.
  • Deploy secure gateway or jump-host architectures to mediate all remote OT access.

2. Purple-Team the OT Environment

Most water utilities operate under a policy framework that includes EPA- and AWIA-mandated risk and resilience assessments and CISA's Cross-Sector Cybersecurity Performance Goals. Others have neither a documented policy nor a means to enforce one, leaving only the baseline regulatory requirement, which may even be loosely followed in practice. The gap exposed in Minnesota, Aliquippa, and Oldsmar was not necessarily a missing policy on paper. It was the absence of a mechanism to confirm whether existing policy, where it existed at all, was actually enforced on the devices and accounts that mattered most. A rule stating "no default credentials on internet-facing PLCs" is only as strong as the last time someone verified it against the live environment. Adversarial engineering teams close that gap by non-destructively testing these controls in isolated or emulated environments:

  • Credential Compromise Simulation: Simulate anomalous authentication to a remote-access account to measure SOC detection and response times.
  • Unauthorized Activity Testing: Emulate an unexpected engineering-workstation project-file push and verify that OT-aware network monitoring alerts on the activity.
  • Trust-Boundary Traversal: In a lab replica, attempt to pivot from a compromised IT asset into the OT network to prove segmentation controls hold.

3. The Ultimate Control: Manual Operations

No network is impenetrable. The definitive saving grace in the Minnesota incidents was the ability to revert to manual operations.

Utilities must formalize and test offline backup restorations and degraded-mode runbooks for every automated control point. Tabletop exercises simulating simultaneous multi-site OT disruptions should be a mandatory annual practice to test manual activation and inter-agency coordination.

Conclusion

The 2026 Minnesota water-utility attacks were not an unpredictable lightning strike of advanced cyber warfare. They were the predictable exploitation of persistent architectural vulnerabilities that have gone unaddressed since at least 2021. Until the sector enforces gateway-mediated segmentation, multi-factor authentication, and robust OT-aware monitoring, threat actors will continue to treat critical infrastructure like a playground until validation catches up to policy.

At Lares, we urge organizations to proactively test their boundaries, assume breach, and ensure that when the screens go dark, the humans can still turn the valves.

[Talk to Lares about a controlled OT/IT boundary assessment →]

Related Article

Minnesota Water Attacks and the Predictable Reality

August 12, 2026 by Andrew Heller How possible Iran-linked actors exploited exposed PLCs in the 2026 Minnesota water attacks. MITRE ATT&CK mapping, risk matrix, and OT security lessons from Lares. Read More Blog, Penetration Testing, Purple Teaming, Red Teaming

Part 3: The Collapse of the Testing Boundary: Why Purple Teaming is Mission-Critical

August 7, 2026 by Andrew Heller Point-in-time pen tests can't catch AI agents that fail in minutes. See why continuous purple teaming is the only defense — and how to build the budget case. Read More Artificial Intelligence, Blog, Purple Teaming

Part 2: The Collapse of the Testing Boundary: Anthropic and the Behavioral Threat

August 7, 2026 by Andrew Heller Claude models attacked real infrastructure while believing they were in a simulation. Anthropic's retrospective reveals a new AI risk class beyond alignment. Read More Artificial Intelligence, Blog, Purple Teaming

Part 1: The Collapse of the Testing Boundary: Deconstructing the OpenAI Agent Breakout

August 7, 2026 by Andrew Heller OpenAI's frontier model escaped its sandbox and breached Hugging Face's cluster in under 13 hours. See the full kill chain mapped to MITRE ATT&CK and ATLAS. Read More Artificial Intelligence, Blog, Purple Teaming

Social Profiling – OSINT for Red/Blue

July 27, 2026 by Lares Labs Read More Blog, Penetration Testing, Red Teaming

The Phantom Menace: Exposing hidden risks through ACLs in Active Directory

June 18, 2026 by Raúl Redondo Discover how attackers exploit hidden risks in Active Directory ACLs. Explore techniques like GenericAll, GenericWrite, and WriteDACL abuse in our latest post. Read More Blog, Insider Threat, Penetration Testing, Red Teaming

Kerberos IV - Delegations

June 17, 2026 by Raúl Redondo Discover how to abuse Kerberos for lateral movement. Learn User Impersonation techniques like Pass the Ticket, Shadow Credentials, and forging tickets. Read More Blog, Blue Team, Penetration Testing, Red Teaming

Kerberos III - User Impersonation

June 17, 2026 by Raúl Redondo Discover how to abuse Kerberos for lateral movement. Learn User Impersonation techniques like Pass the Ticket, Shadow Credentials, and forging tickets. Read More Blog, Blue Team, Penetration Testing, Red Teaming

Kerberos II - Credential Access

June 16, 2026 by Raúl Redondo Dive into the fundamentals of the Kerberos authentication protocol. Explore its history, core concepts, authentication flow, and PKINIT in part one of our series. Read More Blog, Blue Team, Penetration Testing, Red Teaming

Kerberos I - Overview

June 16, 2026 by Raúl Redondo Dive into the fundamentals of the Kerberos authentication protocol. Explore its history, core concepts, authentication flow, and PKINIT in part one of our series. Read More Blog, Blue Team, Penetration Testing, Red Teaming

Empowering Organizations to Maximize Their Security Potential.

Lares is a security consulting firm that helps companies secure electronic, physical, intellectual, and financial assets through a unique blend of assessment, testing, and coaching since 2008.

18+ Years

In business

600+

Customers worldwide

4,500+

Engagements

Where There is Unity, There is Victory

[Ubi concordia, ibi victoria]

– Publius Syrus

Contact Lares Consulting logo (image)

Continuous defensive improvement through adversarial simulation and collaboration.

Email Us

©2025 Lares, a Damovo Company | All rights reserved.