For hospitals, universities, and other critical facilities, that changes the assumptions behind emergency-power planning. Traditional strategies generally anticipate a clear sequence: Utility power fails, generators start, and selected emergency loads remain energized until service returns. That model remains essential, but it may not be enough when outages coincide with extreme temperatures, equipment failures, fuel constraints, or regional grid stress.

The question is no longer simply whether a facility has backup power. It is whether its mechanical and electrical systems can sustain critical operations through increasingly severe and unpredictable conditions.

Extreme Weather is Testing a Grid Under Growing Strain

The North American Electric Reliability Corporation is warning that electricity demand is rising faster than dependable generation resources in several parts of the country. Its 2025 Long-Term Reliability Assessment projects summer peak demand across its assessment areas will grow by 224 GW over the next decade, 69% higher than its 2024 forecast.

Data centers are a major driver, along with industrial expansion, electrification, population growth, and new manufacturing loads. At the same time, generators are retiring and the resource mix is becoming more dependent on weather conditions and the time of day.

Extreme weather is exposing the gaps in the systems.

During Winter Storm Uri in February 2021, record cold increased electricity and heating demand while freezing equipment, generator outages, and natural gas supply disruptions reduced available power. Hospitals then faced a second layer of disruption as water systems lost power or pressure and frozen pipes affected facility operations. The crisis demonstrated that emergency planning cannot consider electricity, fuel, water, and mechanical systems independently.

At the opposite end of the weather spectrum, prolonged heat waves keep air-conditioning demand elevated into the evening, after solar production declines. Thermal power plants may also experience reduced output or performance challenges when ambient temperatures and cooling-water temperatures rise.

Tornadoes, hurricanes, flooding, and severe thunderstorms introduce an additional problem. They may not cause a regional shortage of generation, but they can damage transmission and distribution infrastructure, isolate individual facilities, and delay restoration.

It is not if a facility might deal with these issues, but when.

The Highest Loads May Occur During the Outage

Conventional emergency-power systems are designed to support specific loads and meet essential life-safety requirements. Under National Fire Protection Association mandates, healthcare facilities must plan to sustain operations for 96 hours during an emergency, but backup systems typically serve only designated portions of the electrical infrastructure. If a facility was designed only to the code minimum, much of the building may be without power from the start of an extended outage.

The gap in planning appears when facilities assume those systems can also sustain broader operations through an extended weather-related disruption. Extreme temperatures make that mistake more consequential because the facility may need the most power precisely when the grid becomes least reliable.

During a heat wave, a hospital may have emergency power for clinical equipment but insufficient capacity to maintain chilled water, air movement, and humidity control throughout the facility. A research campus may preserve laboratory freezers while losing the ventilation and exhaust systems required to continue experiments safely. A data center may maintain server power but struggle to support the cooling infrastructure on which those servers depend.

In winter, an electrified building may experience a sharp increase in heating demand. If the emergency-power system was designed before heat pumps, imaging equipment, expanded laboratories, or new campus buildings were added, its original capacity assumptions may no longer reflect actual operations, leaving a significant gap in the backup power systems.

These gaps are especially common on large healthcare and college campuses, as well as in older facilities where systems have expanded incrementally over time. Even some fire and police stations do not have backup power for the entire building. Some universities are addressing the issue by designating specific buildings that can remain fully operational as emergency centers, rather than attempting to support an entire campus.

Resilience planning must account for both mechanical and electrical systems. Teams should identify not only which equipment requires electricity, but also the pumps, controls, cooling, heating, ventilation, fuel, and communications systems on which critical functions depend.

Plan for Compounding Failures

Weather-related outages rarely occur under otherwise normal conditions.

The same storm that interrupts utility service may flood an electrical room, damage cooling equipment, or destroy a road, preventing staff and fuel suppliers from reaching the site. Extreme heat can increase building load while reducing equipment efficiency. Severe cold can affect generators, outdoor equipment, and fuel systems.

Fuel continuity is particularly important. An on-site tank may support operations for a defined period, but a multi-day outage may require replenishment. Owners should know where the fuel will come from, how it will reach the facility if roads are blocked, and whether the supplier will be serving multiple critical customers during the same emergency.

Testing should also go beyond merely starting a generator briefly. Facilities need to evaluate sustained loading, transfer equipment, control sequences, and the loss of individual components. An emergency power system that performs during a scheduled test may respond differently while supporting peak cooling or heating loads for several days.

The planning scenario should not just test what happens if utility power fails. It should analyze what happens if utility power fails during the most demanding weather conditions while at least one supporting component – a pump, a transfer switch, a cooling tower – is simultaneously unavailable.

Reduce Demand Before it Exceeds Capacity

Resilience begins with understanding the loads, not selecting the equipment.

Some systems must remain continuously energized, including life-safety equipment, critical clinical or research functions, controls, communications, and security. Other systems can tolerate a brief interruption, operate at reduced capacity, or be shut down temporarily.

That hierarchy should be translated into an operating sequence through structured load shedding.

During a heat-driven grid stress event, for example, a facility may adjust temperature setpoints, reduce ventilation in unoccupied areas, defer selected processes, or shift loads to on-site resources. If utility power is lost, controls can shed lower-priority loads and preserve available generation for critical operations.

A university might prioritize research laboratories, central communications, and occupied residence halls while reducing cooling in administrative and classroom buildings. A hospital may preserve operating rooms, intensive care, pharmacy storage, and essential environmental controls while postponing nonurgent services.

Structured load shedding is not an energy-efficiency strategy. It is disciplined operational risk management that establishes difficult decisions before an emergency.

Prepare For Long-Duration Islanding

Facilities that cannot tolerate extended disruption may need the ability to operate independently from the utility grid.

Microgrids can provide onsite generation, energy storage, controls, and selected loads. During normal conditions, the system operates in connection with the utility. During an outage, it can disconnect and continue serving supported loads in island mode.

Long-duration islanding requires more than installing a generator or battery. The system must balance changing loads with available generation, maintain voltage and frequency, shed loads when capacity becomes constrained, and resynchronize safely when utility service returns.

Different resources serve different purposes. Batteries can respond quickly, support the transition to island mode, and manage short-term peaks. Solar can reduce daytime demand but cannot sustain operations alone through nighttime or severe weather. Combined heat and power, fuel cells, or conventional generators may provide longer-duration production when supported by a dependable fuel source.

The strongest strategy is often layered. Batteries bridge interruptions and control peaks. On-site generation supports longer operation. Automated controls match available energy to prioritized loads.

These resources can also provide value before an outage occurs. During extreme heat or other events that stress the grid, a facility may use storage, on-site generation, and load shedding to reduce its peak exposure without fully disconnecting from the grid.

Build Resilience Before the Forecast Turns Severe

At Southern Ohio Medical Center, aging and disconnected systems made it difficult to move backup power where it was needed most, putting the hospital at risk in the event of a significant outage. Some generators were already at capacity, while others had available power that could not be shared across the campus. The hospital also relied on several separate chilled-water plants, including older equipment that needed replacing.

The solution centered on integration. A campuswide coordination strategy connected previously isolated emergency generators into a unified backbone, so available capacity could be directed where it was needed most, rather than stranded in one building while another ran short. A new central chilled-water plant replaced aging, fragmented cooling infrastructure and was designed to operate independently of the utility, keeping critical environments stable through an extended outage. For the first time, the electrical and mechanical systems could respond as a single coordinated whole rather than a collection of independent components racing to hold on separately.

The project demonstrates that resilience cannot be divided into isolated mechanical and electrical decisions. Generation resources, central plants, distribution systems, and controls must respond as one coordinated system.

Not every facility needs a full microgrid. Owners can begin by reviewing interval data, identifying weather-sensitive loads, and comparing current demand with the assumptions behind existing emergency systems. Early-stage energy modeling can then test scenarios involving extreme heat, severe cold, equipment failure, fuel constraints, and outages of different durations.

As extreme weather places more strain on both the grid and the buildings it serves, emergency power must become part of a broader operating strategy. The goal is not merely to keep selected equipment running after an outage. It is to reduce demand as conditions worsen, preserve the systems that matter most, and sustain critical operations until both the weather and the grid stabilize.