Microgrids Are Big Tech’s Favorite Workaround—And CRE’s Next Edge

As AI and electrification strain the grid, onsite power is moving from niche experiment to mainstream risk management.

America’s biggest new “tenants” aren’t people or even companies—they’re AI data centers.

AI, Data Centers and a Grid That’s Falling Behind

PJM Interconnection, the largest U.S. grid operator, now expects peak load to grow by 32 gigawatts between 2024 and 2030, with about 94% of that growth coming from data centers. Reuters reporting from this summer noted that PJM has added only about 5 GW of capacity even as demand from AI and cloud facilities pushes capacity auction prices up more than 800%, with some regions bracing for bill increases north of 20%.

At the same time, a mountain of new generation and storage is stuck in interconnection queues. Berkeley Lab’s latest “Queued Up” analysis finds more than 2,500 GW of power capacity—largely wind, solar and storage—waiting for permission to connect, while Canary Media notes that backlog is now roughly twice the size of the existing U.S. grid. Many of those projects will never be built.

That mismatch—load growth measured in months, infrastructure timelines measured in years—is why microgrids have moved from energy conference buzzword to practical workaround.

A recent Reuters analysis estimates U.S. microgrid capacity at 4.4 gigawatts in 2022, on track to reach around 10 GW by the end of 2025, as both Big Tech and utilities look for ways to keep their most critical sites powered during outages and capacity crunches.

Microgrids Go From Niche to Necessary

Microgrids—local energy systems that combine onsite generation, storage and controls, and can island from the grid—used to be framed as boutique resilience projects for hospitals or remote communities. That’s changing quickly.

The Reuters piece highlights how states like California, Texas, Colorado and Georgia are now offering grants or incentives to support microgrid deployment, especially in vulnerable or rural communities. In California, regulators have approved a $200 million Microgrid Incentive Program funded by ratepayers to back community microgrids in disadvantaged and tribal communities that face frequent shutoffs and outages.

Meanwhile, data center developers in PJM and ERCOT are increasingly exploring onsite gas turbines, fuel cells and hybrid renewable-plus-storage systems to avoid multi-year waits for new transmission, sometimes pairing these with behind-the-meter microgrid controls to keep loads within contracted limits and ride through grid stress events. Texas, facing its own rapid build-out, has gone as far as passing a law allowing the grid operator to curtail or disconnect data centers and other large users during emergencies, and another that requires very large loads to meet specific interconnection standards and share in transmission costs.

For commercial and institutional property owners, the signal is clear: onsite power is becoming standard practice wherever load growth and reliability are strategic risks, not just operational nuisances.

From “Bridge Power” to Campus Operating Systems

In our own coverage at ChargedUp!, we’ve described one emerging model as “bridge power”—modular systems designed to energize a site in months, not years, and then keep earning their keep after the permanent grid upgrades arrive. DG Matrix, for example, is positioning its solid-state “power router” as the heart of such systems, capable of blending multiple sources—grid, renewables, battery storage, fuel cells and backup generators—into one harmonized architecture. In practice, that means a developer can use pre-engineered, containerized components to get a campus, depot or data facility live quickly, then repurpose those same assets for peak shaving, backup and demand-response revenue once a full interconnection is in place.

The appeal is speed and optionality. Bridge-style microgrids can compress time-to-energize, a critical metric when a stalled power connection can derail a business plan. They can also be redeployed or reconfigured across sites, reducing the risk of “stranded” generators and switchgear that no longer fit the load profile.

There’s a governance story here, too. In West Virginia, lawmakers recently passed the Power Generation and Consumption Act, designed in part to attract high-impact data centers and “small energy grids.” The law stripped local governments of most authority over these projects and redirected a large share of tax revenue to the state, prompting backlash from residents who discovered plans for a gas-fired data center microgrid only after permits had been filed.

Other states are debating how far to go in giving large private loads special rules—or special curtailment obligations—during grid stress. For landlords and campus owners, that mix of incentives, carve-outs and local pushback is the new backdrop for major on-site power decisions.

Questions for Owners and Operators to Ask

Before you green-light a microgrid or “bridge power” project, it’s worth pressing on a few specifics:

• What problem are we actually solving? Is the primary driver speed-to-energize (avoiding an interconnection delay), resilience (keeping critical loads up during outages), tariff management (demand charges and time-of-use), or some combination—and how does that tie back to NOI, uptime or mission?

  
  

• How will this evolve with our campus plan? Can the system scale to future loads—EV charging, labs, robotics, additional buildings—or is it sized for a single tenant or use case that might change in five years?

  
  

• Where do regulators and communities fit? Are we in a jurisdiction that’s actively encouraging microgrids for resilience, or one where local land-use and air-quality rules could become friction points, especially if thermal generation is involved?

  
  

• How do we avoid stranded or “privatized” infrastructure? If tariffs change or a major tenant leaves, can assets be reconfigured, expanded or redeployed, or are we locked into a footprint that only makes sense under today’s assumptions?

  
  

• Who controls the software and data? Microgrids increasingly depend on sophisticated controls that integrate with building management systems, DERMS and EV charging networks. Do you retain clear rights to operational data, APIs and future integrations, or are you locked into one vendor’s black box?

  
  

• What’s the finance-ready story? Can you show lenders and underwriters—in a simple model—how the project reduces specific risks (outages, delay penalties, demand charges) and supports long-term asset value, rather than just adding complexity and capex?

A New Kind of Location Advantage

For the next decade, power will be as strategic to asset performance as location and lease structure. Hyperscalers are already acting on that premise, building microgrids and “bridge power” systems as default infrastructure where the grid can’t keep up. State programs like California’s $200 million Microgrid Incentive Program, and similar efforts in Texas and elsewhere, show policymakers are starting to see microgrids as a public resilience tool as well as a private hedge.

For commercial, industrial and campus owners, the opportunity is to adapt those playbooks carefully—treating on-site power not as a gadget, but as part of an intelligent electrification stack that also includes EV charging, storage, connectivity and tenant-facing services. The risk is that, in the rush to solve today’s capacity crunch, projects lock in narrow benefits and broad liabilities.

The microgrid moment is here. The portfolios that benefit most will be the ones that ask sharper questions—and insist that their new power systems serve both the balance sheet and the communities around them.