The United States today runs on economic lifelines that Americans take for granted: electricity, clean water, food, information technology, communications, transportation, fuel, and hospitals. But they are all dependent on a power grid that was not designed for demands like cascading high-impact disasters and the unexpected growth of new data centers, such as those required by artificial intelligence. According to a study by the U.S. Department of Energy, up to 108 GW of power may be needed just for data centers.

These concerns are heightened even more now as emerging energy demand strains power grids for what many consider the next arms and economic race: AI and its data centers. Billions of dollars are being spent by AI investors on the construction of data centers now, including those driven by AI needs, often without committed power sources. Their financial and operational success will depend on the availability of that power.

Yet a new optimism is emerging among planners, entrepreneurs, and defense leaders who recognize that the same interdependence that creates vulnerability could also build resilience. But that resilience will only be achieved if the needed energy systems are generated and managed locally. Some emerging data centers may fund those systems in the interest of their own success and possibly their survival.

From Centralized Strength to Distributed Survival, Local Energy is Essential

The U.S. electric grid – one of the largest machines in history – delivers extraordinary efficiency. As the grid developed over time, a two-fold shift happened. When electric power that once was merely a modern convenience to consumers and an economic advantage to industry became essential to lifeline infrastructures, efficiency was relentlessly pursued, and regional grids maximized cost-effective, centralized control.

In a July 7, 2025, report, the Department of Energy makes it clear that, with retiring units and the highest load growth since the 1970s, the power system is becoming less resilient. Like other critical agencies, all 16 of the Department of Homeland Security’s critical infrastructures are dependent on power. A whole new reality is emerging: Efficiency has come at the expense of resilience due to over-centralized systems. Cascading blackouts caused by storms, cyber incidents, or supply disruptions ripple across larger regions because generation and transmission remain tightly interconnected (see Figure 1), unable to operate in island mode.

Since 2000, special attention has been focused on nationwide, high-impact threats that could disable power grids for 1–18 months, such as cyberattacks, coordinated physical attacks, or electromagnetic pulse (EMP) events. (Caused by natural phenomena such as extreme space weather or by foreign adversaries, EMPs can cause widespread, long-term power outages.) This attention has resulted in reports by the U.S. Congressional EMP Commission in 2004 and 2008 (Executive Report; Full Report) and the 2010 North American Electric Reliability Corporation (NERC) report High-Impact, Low-Frequency Event Risk to the North American Bulk Power System as well as the Defense Threat Reduction Agency (DTRA) public call for EMP-protected microgrids in its 2015 Small Business Innovation Research request for proposals. In that request, DTRA explained that U.S. military bases are dependent on commercial power grids that are not EMP-protected and, therefore, could be unable to receive power “permanently, or for weeks or months.”

The Need and Demand for Island-Mode Microgrids

These concerns led to DTRA’s 2015 call for EMP-protected microgrids that could conduct critical missions in “island mode” for months or years, that is, operating from local power sources independent of a wider grid. This expectation requires a new class of local energy systems to withstand physical, cyber, and electromagnetic threats that can cripple centralized systems. In effect, microgrids must draw on anything overhead, underfoot, or nearby – solar and wind with batteries on site, geothermal or micro nuclear underground, and locally available fuels – rather than distant sources that may become inaccessible. Similar requirements appear in hospital energy standards set by the Joint Commission, ensuring medical facilities can stay open when they are needed most.

Taking into account emerging data centers, a lack of sufficient power can delay their market entry and reduce their returns on investment – potentially halting some projects. Figure 2 shows how on-site energy systems can reduce the risk of those investments and mean the difference between success and failure.

In earlier decades, large hydro, fossil, and nuclear plants dominated power production but depended on complex supply chains and long transmission corridors. Large utility systems could leverage large-scale capabilities to produce power less expensively than smaller local systems. However, energy providers and users who need to be able to operate in island mode are shifting toward a hybrid approach: Regional electric utilities operate their systems but also connect with on-site distributed energy resources, including energy storage systems that provide the flexibility to operate through a long-term widespread power outage. This helps them isolate, adapt, and recover from power outages, and in many cases, this helps larger grids restart.

Emerging Technologies in Local Generation Worth Watching

New classes of energy technology are challenging traditional assumptions about scale and feasibility. Innovations that are at development stages classified as technology readiness levels 4–8 are now entering early commercialization stages:

• 10-megawatt Micro-Reactors – Compact and safe reactors capable of providing modular multi-year baseload power to communities, military bases, or industrial clusters with minimal refueling.
• Atmospheric (Ion Collection) Energy Systems – Technologies harvesting electrostatic potential from the atmosphere to generate low-current but high-voltage, continuous power for communications, sensors, and limited basic needs.
• Hydrogen Generation and Storage – Modular systems producing “green” hydrogen on site with long-duration storage and to generate usable fuel on site, including jet fuel.
• Refined Geothermal, Hydro, Solar, and Wind Hybrids – Systems integrating multiple renewable sources into cohesive, islanded microgrids.

At different maturity levels, each of these technologies, when in use, outline a near-term future in which local, modular systems can operate effectively in island mode and complement centralized systems that are less flexible and more brittle on their own.

Lessons From Chicopee and Westover Air Reserve Base

On September 10, 2024, the City of Chicopee, Massachusetts, and Westover Air Reserve Base hosted a landmark exercise: Imagining the U.S. Without Power – A Dual-World EMP Tabletop Exercise. (The article includes the link to the final after-action report.) The scenario contrasted two “worlds”:

• Group A, operating with no new resilience measures, and
• Group B, operating with limited EMP-protected microgrids, one-year food reserves, and hardened communications.

Near the end of the exercise, each group separately discussed what they had learned in a debrief. Then both groups were brought together to compare their respective takeaways. The differences were dramatic. In the exercise scenario, Group A saw a 50–75% population loss after six months without grid power and loss of continuity of government. Group B, with just 20–30% power, maintained rice-and-bean reserves for a third of the population and retained social order, continuity of government, and survival rates exceeding 70%. (Projected nationally, that approximately 25% difference in lives lost could account for 83.75 million people. It is worth noting that even in the better-case scenario of Group B, over 100 million people still die across the country, since scenario B only assumed partial preparation, as opposed to none. In contrast, 53,000,000 lives were lost in all of World War 2.)

The key insight from this exercise was that even modest preparation can save lives and preserve continuity of government. Planning, distributed power, food, and communications can prevent panic after an EMP event or any other cause knocking out centralized power across the country for extended periods.

The National Security Dividend

The Chicopee–Westover exercise revealed a sobering reality: Without local resilience, the Department of War cannot project the available force. In a prolonged grid outage, the Chicopee-Westover participants expressed concern that service members whose families are suffering and dying of starvation will be less ready to deploy when needed. That recognition reframed microgrids as strategic assets, not luxury redundancies.

These resilient island mode–capable microgrids could work as dual-use energy systems that support national defense and civilian stability. Bases, hospitals, and local governments that host microgrids can sustain their missions while supporting neighboring populations. The DTRA-supported EMP microgrid initiative and Resilient Hospitals Handbook, funded by the U.S. Department of Homeland Security, demonstrate how such systems become “lifeboats” during cascading failures. They also show that these island mode–capable microgrids can and must be protected from the same cyber, physical, and electromagnetic threats facing larger commercial grids.

Financing the Shift

Transitioning from fragile dependence to distributed resilience demands financial creativity. Fortunately, some policy tools have been initiated, but it is unclear what will be changed, dropped, or added:

• Department of Defense’s Defense Critical Infrastructure Program funds energy efforts at bases and surrounding communities.
• Federal Emergency Management Agency’s efforts support microgrid integration for emergency services. (Due to imminent changes to the agency’s funding, communities should seek private investment by making the best business case possible.)
• Department of Energy programs promote grid modernization and cybersecurity integration.
• “Resilience-” or “Energy-as-a-Service” project financing models optimize public and private capital blends, delivering critical infrastructure and predictable returns on investment at relatively low upfront capital costs.

Resilience-as-a-service – such as power purchasing agreements that attract third-party financing to fully pay for upfront costs of a microgrid – leverages the owner’s cost savings and revenue with whatever incentives come and go. This helps ensure that private-sector investors – especially AI data-center investors – can achieve a return on investment while sharing their local energy with neighbors. The same distributed systems that guarantee data uptime during commercial outages also protect local water, food, and healthcare infrastructure that the data centers also need (see Figure 3).

Managing Skepticism – From Fragile Grid to Living Network

Preparedness communication often swings between two extremes: complacency or catastrophe. Both lead to paralysis. Some professionals describe a new emotional challenge – “pre-traumatic stress disorder” – by which the fear of inevitable collapse discourages constructive action. Resilience planning must reject that mindset. Each new microgrid, each localized supply chain, and each trained technician is a step toward hope, grounded in design and multiple opportunities for return on investment. This work is neither speculative nor alarmist – it provides the architecture of confidence.

The creation of the U.S. electric grid was determined to be the most important engineering achievement of the last century by the National Academy of Engineering. It is entering the most transformative era of its evolution. Centralized assets will remain essential and helpful for the financial success of local energy systems, but survivability will depend on local and islanded, modular networks that act as both energy producers and community stabilizers. The vision emerging from the Chicopee–Westover Air Reserve Base tabletop exercise and professional conferences that bring together microgrid users and energy providers, such as the Energy and Mobility Conference, is clear: Distributed resilience is no longer a technical experiment – it is an economic and moral imperative. In light of the DTRA example that declares that military bases and organizations that support them need EMP- and cyber-protected microgrids, local communities could work with the well-funded investors of new data centers to create resilient local power generation. This way the financial viability and security of both data centers and local communities can be established before an energy disaster, so that they can maintain resilience and power throughout a long-term grid failure. Whether for defense, data, or daily life, those who invest in local power and microgrids will see a return in national grid strength and the resilience of their own local infrastructure projects, such as data centers.

Take Action

In summary, entrepreneurs, data centers or electric utilities owners and operators, government officials, and regulators can consider these actions to promote local energy independence.

• Be financially creative to reduce grid interdependencies.
• Protect microgrids from EMPs to ensure continuity of critical missions in “island mode” for months or years.
• Consider shifting energy providers and users to a hybrid approach to facilitate operations in island mode.
• Seek out energy conferences that facilitate the meeting of power technology providers and data center operators.

Acknowledgments: Special thanks to the Westover–Chicopee EMP Tabletop Exercise participants, the Foundation for Infrastructure Resilience, the International Council on Systems Engineering (INCOSE) and its Critical Infrastructure Protection and Resilience Working Group, Cleveland State University, and the numerous entrepreneurs and investors advancing emerging microgrids that strengthen both community and national resilience.