ARTICLE: Why Data Centers & Large Facilities Should Pay Attention To NERC’s Level 3 Alert
NERC issued a Level 3 Essential Actions Alert in May 2026 after a series of large-load reduction and oscillation events raised concerns about bulk power system reliability. NERC alerts are industry notifications used to identify emerging reliability risks and request action, information or acknowledgement from registered grid entities.
This alert is formally aimed at transmission planners, transmission owners and operators, balancing authorities, reliability coordinators and other registered entities. The practical effect, however, will be felt across the wider energy ecosystem because those entities depend on accurate facility data, operating models, commissioning evidence and real-time coordination from the loads and distributed resources connected to their networks.
Data centers are a key segment to watch. AI and high-performance computing sites can concentrate hundreds of megawatts of demand behind complex electrical systems, UPS architectures, backup generation, controls, cooling loads and staged buildouts. Their grid impact is increasingly defined not only by how much energy they consume, but by how fast they can reduce, transfer, restart or change operating state.
Why data centers are central to the alert
The concern behind the alert is speed and scale. A large facility that suddenly drops load, transfers to backup supply, changes UPS mode or reconnects after a disturbance can materially affect voltage, frequency and system balance. If several sites or connected resources behave in similar ways, those changes can happen faster than operators can study, coordinate or respond in real time.
Data centers sit at the center of this issue because computational load can be both large and operationally dynamic. Load profiles can change with IT utilisation, cooling demand, backup-system behaviour, curtailment programs, on-site generation, storage and control logic. For utilities and grid operators, conventional assumptions about steady industrial load may no longer be sufficient.
What changes for the energy ecosystem
Modelling expectations will rise. Utilities and transmission entities will require clearer information on computational load, non-IT load, UPS behaviour, backup generation, storage, power factor, ramp rates, phased energisation and protection settings.
Net-load behaviour will matter more. Facilities will need to explain when they consume, reduce, transfer, export, island, reconnect or restart, and what controls or protection systems trigger those actions.
Change notification will become more important. Load growth, new IT blocks, UPS mode changes, protection-setting changes, DER additions and facility repurposing may need to be communicated earlier than traditional planning cycles allow.
Operating coordination will tighten. Expect more emphasis on named contacts, event data, SCADA points, commissioning evidence, emergency procedures and coordination with utilities before and during grid events.
Actions for Facility teams
1. Create a utility-ready technical data package
Prepare a shareable package covering site one-line diagrams, maximum and minimum MW, expected peak demand, phased buildout schedule, IT versus non-IT load, cooling load, power factor, UPS behaviour, generator modes, storage operation, protection settings and relevant SCADA points. The package should be controlled, current and easy to update as the facility evolves.
2. Document ramp-down, transfer and restart behaviour
Define how quickly the site can reduce load, transfer to backup supply, reconnect, restart IT load and return cooling systems to normal operation. Include the events that trigger each action, whether the response is automated or manual, and the conditions required before the site returns to grid supply.
3. Coordinate studies before changes become operational
Utilities may need to study full-load operation, partial-load operation, UPS transfer, backup generation transfer, emergency curtailment, storage dispatch, reconnection and staged energisation. Data center owners and operators should engage early, especially when adding major IT blocks, changing electrical architecture or introducing behind-the-meter resources.
4. Strengthen commissioning and event monitoring
Commissioning should validate the model that the utility is relying on. Test protection settings, SCADA points, breaker status, UPS modes, generator transfer, storage behaviour and operating communications. Where feasible, capture disturbance data at the point of common coupling so post-event analysis is based on evidence rather than assumptions.
5. Put vendor and operator controls around reliability-critical settings
OEMs, controls vendors, UPS suppliers, cooling providers and site operators should be required to provide settings data, model data, firmware-change notifications and test support. Changes that affect load behaviour, trip behaviour, restart timing or export behaviour should follow a formal approval and notification process.
The strategic shift
The alert does not turn data centers into registered grid entities. It does, however, make clear that large, fast-moving facilities are no longer passive customers from a reliability perspective. Their behaviour has become part of system planning, system operations and grid-risk management.
For the broader ecosystem of facilities, utilities, distributed generation, storage and aggregators, grid reliability will depend on better data, better models and more disciplined coordination between operators and connected assets. Data centers are the most visible near-term example, but the same logic will increasingly apply to any site or portfolio whose load or generation can move quickly enough to affect the grid.
How AZZO can help
AZZO supports complex energy users and distributed energy stakeholders with the data, controls and operational visibility required to understand and explain site behaviour. For data centers and other grid-interactive facilities, this means clearer modelling inputs, better monitoring, stronger change management and practical coordination with utilities and network operators.
As reliability expectations evolve, organisations that can model, monitor and control their energy behaviour will be better positioned to connect, operate, expand and participate in a more dynamic grid.