FERC Moves to Bring Data Centers and Other Computational Loads Into the Mandatory Reliability Framework

By Ryan J. Regula and Sukhmani K. Singh

The Federal Energy Regulatory Commission (FERC or the Commission) has taken a significant step toward bringing data centers, cryptocurrency mining, artificial intelligence infrastructure, and other intensive information technology loads into the mandatory reliability framework.

On July 16, 2026, the Commission directed the North American Electric Reliability Corporation (NERC) to file new or modified Reliability Standards by December 31, 2026, addressing risks associated with integrating “computational loads” into the Bulk-Power System. FERC also directed NERC to develop related Rules of Procedure revisions, including registry criteria for computational load entities, by the same year-end deadline, and to submit a March 1, 2027, informational filing describing the next phase of standards development. The practical message is clear: a large load strategy can no longer be treated only as an interconnection, siting, power procurement, or commercial contracting issue. For many projects, it is becoming a reliability compliance issue as well.

Why It Matters

FERC acted under Section 215(d)(5) of the Federal Power Act, which authorizes the Commission to direct the Electric Reliability Organization to submit a proposed Reliability Standard or modification addressing a specific reliability matter. Once approved, Reliability Standards are mandatory and enforceable in the United States by NERC, subject to FERC oversight, or by FERC independently. That makes the order more than a request for best practices. It starts a process that may impose enforceable obligations on entities whose load characteristics affect Bulk-Power System reliability.

The order also fits within a broader federal focus on large load interconnection. In October 2025, the Secretary of Energy directed FERC to consider reforms for timely and orderly interconnection of large loads to the interstate transmission system. FERC’s large load proceeding seeks input on potential reforms for loads generally greater than 20 megawatts, including data centers and manufacturing facilities. NERC told FERC that demand growth is higher than at any point in the last two decades and that large loads are being built faster and at larger scale than prior loads using the Bulk-Power System. FERC accepted the need for prompt action, citing unprecedented load growth driven by data centers and NERC-documented disturbances in which computational loads caused or contributed to Bulk-Power System instability.

Who Could Be Affected

The order itself does not make every data center, host, customer, utility, or developer subject to new Reliability Standard obligations. Instead, it directs NERC to develop the standards, definitions, and registration framework that may later determine who is responsible for compliance. NERC’s current proposed Rules of Procedure revisions would define “Computational Load” as electric power demand from information technology equipment, such as servers, storage, and networking hardware. They would define a “Computational Load Entity” as the end user or entity that hosts end users receiving electric power for Computational Load.

NERC’s April 2026 proposal indicates that registration could apply where an entity contributes to aggregate connected load capability of at least 20 megawatts, connects at a single Bulk-Power System point of interconnection at 60 kilovolts or above, and hosts at least 1 megawatt of Computational Load. Those criteria remain subject to the NERC and FERC approval processes, but they are already important planning markers for projects in development. The affected universe may include owners, tenants, colocation providers, hosts, utilities, transmission owners, load-serving entities, and project sponsors that must determine which legal entity will bear registration and compliance responsibility.

What Companies Should Consider Doing Now

Companies developing, hosting, financing, or serving computational loads should start with a threshold assessment. That assessment should identify aggregate connected load capability, point of interconnection, interconnection voltage, power demand attributable to information technology equipment, and the entity or entities receiving or hosting power for the load. It should also consider phased campuses, expansions, adjacent sites, and tenant structures that could affect aggregate impact.

Companies should also review interconnection, utility service, colocation, customer, procurement, and financing documents for reliability compliance allocation. Many contracts may not clearly allocate responsibility for data submissions, modeling assumptions, operational limitations, telemetry, ride-through capabilities, curtailment performance, compliance documentation, or future compliance costs. If a host is registered but a tenant controls equipment or operating practices, the host may need contractual rights to obtain information and enforce technical requirements.

Technical and legal teams should coordinate before standards are final. NERC has emphasized that computational loads may have distinctive and less predictable electrical behavior, and those characteristics are central to the proposed Computational Load Entity framework. Equipment configuration, protection settings, backup systems, power electronics, controls, staged energization, and load-shedding functionality may therefore have compliance consequences.

The Bottom Line

FERC’s order is a turning point for computational load regulation. It does not immediately register every data center with NERC, but it directs NERC to build the standards and registration framework that could bring certain computational load entities into the mandatory reliability regime. The December 31, 2026, deadline means the first wave of obligations may arrive quickly relative to development, financing, procurement, and interconnection timelines. Companies that treat the order as a present planning issue may be better positioned to manage regulatory risk, preserve development flexibility, and avoid disputes over who owns the operational and legal consequences of computational load integration.