Energy and water utilities are often asked to show a continuity plan, a cyber program, and a vendor list. The harder question is whether a utility can keep essential service flowing when a supplier, a control-system vendor, or a model used in operations fails. For many organizations those three issues sit under different managers and different regulatory labels, even though they describe the same operating problem.
The official landscape is split, and it is a mistake to blur the split. Applicable users, owners, and operators of the Bulk Electric System are subject to mandatory NERC Critical Infrastructure Protection reliability standards, which the Federal Energy Regulatory Commission approves and which include supply-chain and recovery obligations. Community drinking-water systems that serve more than 3,300 people are subject to America's Water Infrastructure Act section 2013, which amended section 1433 of the Safe Drinking Water Act and requires a risk and resilience assessment and an emergency response plan, recertified on a five-year cycle. Wastewater systems are encouraged to plan; they are not required to certify those products to EPA. Public power may fall under CIP, AWIA, both, or neither, depending on registration and on whether the organization also operates a community water system.
This article treats that split as the starting point, then asks how third-party risk and AI used in operations should be governed so that recovery, supply-chain control, and resilience assessments describe the same set of critical dependencies.
Why this matters
A public-power utility, a joint-action agency, or an investor-owned utility can depend on a small number of vendors for energy-management systems, outage management, customer information, and the industrial control systems that operate substations and generating units. A water or wastewater system can depend on just as few vendors for supervisory control, treatment-process control, laboratory information, and chemical supply. In both cases, a supplier interruption or a compromised update can become a service interruption.
FERC's authority to approve mandatory cybersecurity reliability standards comes from the Energy Policy Act of 2005. FERC has certified the North American Electric Reliability Corporation as the Electric Reliability Organization. On January 18, 2008, FERC issued Order No. 706, approving CIP reliability standards while directing further work. The standards have been revised since then, but the structure remains: once FERC approves a CIP standard, it is mandatory and enforceable for applicable BES entities. Not every municipal or cooperative utility is a registered BES entity, and CIP should not be described as a general "utilities cyber rule." Where it applies, it is not optional.
Water systems have a different statutory clock. Under SDWA section 1433, as amended by AWIA section 2013, community water systems serving more than 3,300 people must develop or update a risk and resilience assessment and an emergency response plan, and certify completion to EPA. Assessments must consider malevolent acts and natural hazards; the resilience of physical and electronic systems, including the security of automated systems; monitoring; financial infrastructure; chemical use and storage; and operations and maintenance. The emergency response plan, due no later than six months after the assessment is certified, must include strategies and resources to improve resilience, including physical security and cybersecurity; procedures and equipment that can be used when a hazard threatens the delivery of safe drinking water; actions that can lessen the public-health impact, including alternative source-water options where relevant; and strategies to aid detection of threats. The next five-year certification deadlines for systems serving 3,301 to 49,999 people are June 30, 2026, for the assessment and December 31, 2026, for the emergency response plan if the assessment was certified on the final due date. That second date is close. EPA does not require the use of any particular third-party standard or tool; the system remains responsible for meeting the statutory elements.
Supply chain and recovery under NERC CIP
Two CIP standards are especially relevant to third-party risk and operational resilience. CIP-013-2, Cyber Security — Supply Chain Risk Management, is mandatory and subject to enforcement as of October 1, 2022. Its stated purpose is to mitigate cyber security risks to the reliable operation of the BES by implementing security controls for supply-chain risk management of BES Cyber Systems. NERC's official record lists a FERC order date of March 18, 2021, an effective date of October 1, 2022, and an inactive date of June 30, 2028, with CIP-013-3 listed for future enforcement beginning July 1, 2028. Until that later date, CIP-013-2 is the enforceable version.
CIP-009-6, Cyber Security — Recovery Plans for BES Cyber Systems, is mandatory and subject to enforcement as of July 1, 2016. Its stated purpose is to recover reliability functions performed by BES Cyber Systems by specifying recovery-plan requirements in support of the continued stability, operability, and reliability of the BES. NERC likewise lists an inactive date of June 30, 2028, for this version. The pairing is the point. A supply-chain control that never informs the recovery plan, or a recovery plan that assumes a vendor will be available without having been part of the supply-chain process, is not a complete reading of the two standards.
CIP obligations are entity- and system-specific. Applicable entities should work from the current enforceable standard, their registration, and the BES Cyber Systems they have identified, not from a generic "CIP program" description. This article is not a compliance checklist and is not an official interpretation of any CIP standard.
Water, wastewater, and public power
The AWIA/SDWA section 1433 duties are a resilience and planning obligation, not a CIP analogue. They do, however, force a utility to look at electronic and automated systems, cybersecurity, and the ability to deliver safe drinking water after a malevolent act or a natural hazard. Vendor-hosted control systems, remote support connections, and cloud-based customer or laboratory platforms belong in that picture if they affect those statutory elements. So do chemical and equipment suppliers, because the assessment must address chemical use, storage, and handling, and the plan must identify equipment that can be used when a hazard threatens delivery.
Wastewater systems, very small community systems, and non-community systems are not required to certify an assessment or plan to EPA under section 1433. EPA encourages those systems to conduct assessments and develop plans anyway. A wastewater utility that shares vendors, networks, or staff with a sister water system should not assume that the water system's certification covers the wastewater operation.
Public-power utilities illustrate why labels are not enough. A municipal electric utility may be a registered BES entity with CIP-013 and CIP-009 obligations, a smaller distribution system outside those standards, a water-system operator with an approaching SDWA certification date, or some combination of those roles. The board or governing body still needs a single view of which services cannot stop, which suppliers and models those services depend on, and which official obligation, if any, already requires that view to be certified or audited.
AI on operational technology is still a third-party question
Utilities are exploring AI for load and outage forecasting, asset inspection, anomaly detection, treatment-process optimization, and customer operations. Some of those tools sit on corporate networks. Some sit close to operational technology. Many are trained, hosted, or updated by vendors who also supply the underlying platform. A model that recommends a switching action, flags a cyber anomaly, or adjusts a process parameter is not only an innovation project. It is a potential high-impact dependency, and it may be a supply-chain dependency under CIP-013 where BES Cyber Systems are involved.
The National Institute of Standards and Technology published the Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1, on January 26, 2023. It is voluntary and organized around Govern, Map, Measure, and Manage. It describes trustworthy AI as valid and reliable, safe, secure and resilient, accountable and transparent, explainable and interpretable, privacy-enhanced, and fair, with harmful bias managed. "Secure and resilient" and "safe" are the characteristics a utility board should hear first when a model is proposed for operations.
ISO/IEC 42001:2023, published in December 2023, provides a management-system standard for organizations that provide or use AI. Certification is voluntary and is not performed by ISO. Used carefully, it can bring vendor-supplied and internally developed models under the same policies, roles, and lifecycle controls, including change management and intended-use definition.
Organizations that place or deploy AI in the European Union also need to account for Regulation (EU) 2024/1689, the EU Artificial Intelligence Act. The Act entered into force on 1 August 2024 and applies in phases; most remaining obligations have applied since 2 August 2026, with some high-risk AI rules applying later. The Act treats certain uses, including some in critical infrastructure, as high-risk, and some of those rules apply later than 2 August 2026. Whether a particular utility use is in scope is a legal and factual question. The operational question is available now: what is the model's intended use, who can change it, how is it tested safely, and how is it disabled if it misbehaves or is unavailable?
Where the three issues meet
The meeting point is a reliability or public-health function that cannot pause: the ability to operate or restore BES Cyber Systems, to treat and deliver safe drinking water, or to maintain control of a wastewater process. The vendor's delayed patch, the unavailable spare, or the withdrawn model becomes the utility's service problem. CIP-013 and CIP-009 already require supply-chain and recovery discipline for applicable BES Cyber Systems. SDWA section 1433 already requires certified assessments and plans that include automated systems and cybersecurity for covered water systems. AI does not sit outside those duties because it is new.
Exercises and recovery tests should therefore include supplier and model failure, not only weather and physical damage. Contracts and procurement should address secure development and update practices, incident notification, spare equipment or alternative sources, and the right to information needed for recovery. Maps of IT and operational-technology dependencies should include the vendors and models in the path, not only the owned assets. See also Mapping Dependencies Before Disruption, Beyond Due Diligence: Proving Third-Party Resilience, and The Hidden Risk of Concentration.
What organizations should do
Separate the official obligations first. If you are an applicable BES entity, work from the current enforceable CIP-013 and CIP-009 standards and the systems they cover. If you operate a community water system serving more than 3,300 people, work from the SDWA section 1433 elements and the certification cycle that applies to your population band, including the December 31, 2026, emergency-response-plan date for systems that certified an assessment on June 30, 2026. Do not use CIP language to describe an AWIA duty, or the reverse.
Build one inventory of critical operational functions, the vendors that support them, and any AI used in or near those functions. Use that inventory in the CIP supply-chain and recovery processes, in the SDWA assessment and plan, or in both, as applicable.
Put recovery evidence next to supply-chain evidence. Know how you would operate or restore a critical function if a vendor cannot perform or if a model must be taken out of service. Test that assumption.
Govern operational AI with a documented intended use, human oversight, a safe testing path, monitoring, and a discontinuation plan. Use the NIST AI RMF and, where a management-system frame is useful, ISO/IEC 42001. Do not claim conformity you have not established, and do not place a model in a control path without a fail-safe that operators understand.
Give the governing body a combined report, not three unconnected packets. For a broader discussion of evidence leaders can challenge, see From Plans to Proven Readiness and Operational Resilience: Where to Start.
How OpResONE can help
OpResONE helps electric, public-power, and water utilities strengthen operational resilience, supply-chain oversight, and AI governance as a single management problem, including dependency mapping and recovery-test design. Our work is independent. We have no affiliation with FERC, NERC, EPA, NIST, or ISO, and this article is not an official interpretation of their standards or statutes. OpResONE does not perform AWIA certifications.
We bring BRILLIANCE to RESILIENCE®.
Related resources
- Mapping Dependencies Before Disruption
- Beyond Due Diligence: Proving Third-Party Resilience
- The Hidden Risk of Concentration
- When a Critical Provider Cannot Recover
- Connecting AI Governance and Resilience
- From Plans to Proven Readiness
- Operational Resilience: Where to Start
Request a Strategy Call
If you want a focused reading of how supply-chain, recovery, and AI oversight line up with the official resilience duties that apply to your electric, water, or wastewater operations, talk with OpResONE about where to start.
Sources
- FERC, Cyber and Grid Security (Energy Policy Act authority; Order No. 706, January 18, 2008)
- NERC CIP-013-2, Cyber Security — Supply Chain Risk Management (mandatory October 1, 2022)
- NERC CIP-013-3, Cyber Security — Supply Chain Risk Management (future enforcement July 1, 2028)
- NERC CIP-009-6, Cyber Security — Recovery Plans for BES Cyber Systems (mandatory July 1, 2016)
- EPA, AWIA Section 2013 / SDWA Section 1433: Risk and Resilience Assessments and Emergency Response Plans
- NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1 (January 26, 2023)
- ISO/IEC 42001:2023, Information technology — Artificial intelligence — Management system
- Regulation (EU) 2024/1689 (EU Artificial Intelligence Act), EUR-Lex
- European Commission, AI Act overview

