Data Centers, Solar, Storage, and Microgrids Are Converging into a Behind-the-Meter Playbook
By Keith Reynolds | Publisher & Editor, ChargedUp!
For years, behind-the-meter energy projects lived in a familiar box: solar for ESG, batteries for demand charges, generators for emergencies. Hyperscale data centers are blowing up that taxonomy. When your load jumps from the equivalent required to power a large building to the load required to run a small small power plant, energy stops being an operating expense and becomes a way to lock in megawatts on construction timelines that the grid can’t reliably meet.
That logic is now spilling into adjacent commercial real estate. Logistics campuses, industrial parks, and large parking assets share feeders and substations with AI-driven load clusters. When the biggest neighbor on your circuit starts buying resilience and flexibility, everyone else feels the ripple in interconnection timelines, upgrade costs, and tariff risk.
The energy park response: build power alongside compute
One of the clearest signals is how quickly hyperscalers and data center developers have shifted from procuring clean power to procuring power. Wood Mackenzie, via pv magazine USA, describes developers co-locating solar and battery storage to avoid interconnection queues, using configurations like private wire and direct connect so new generation can serve new load without waiting years for traditional grid upgrades. This activity a new project delivery model: phase the data center, phase the generation, and use modular solar and storage to keep pace with load growth.
Reliability math is pushing the same direction
This isn’t happening in a vacuum. Grid planners are increasingly blunt about how fast the load picture is changing. PJM Interconnection’s 2026 load forecast projects summer peak load growth averaging 3.6% per year over the next decade, with summer peak forecast around 222,106 MW in 2036 (a ~65,733 MW 10-year increase).
The latest Long-Term Reliability Assessment (LTRA) from the North American Electric Reliability Corporation (NERC) frames the data center boom as a major variable in load forecasting and reliability planning. The report highlights how regions like the mid-Atlantic (served by PJM) are seeing large-load activity that is both massive and difficult to time precisely.
Utility Dive’s coverage of the NERC assessment underscores the scale of the shift. NERC forecasts 224 GW of summer peak demand growth over the next 10 years, with new data centers accounting for most of the projected increase.
These sources suggest that even if some projects slip, the direction is clear enough that the best-capitalized developers are no longer waiting for perfect certainty. They’re buying optionality, often behind the meter.
Batteries become the default firming layer
As soon as you treat power delivery as schedule risk, batteries stop being a finance-side add-on and become a construction-side enabler. Reuters reports global battery storage installations jumped 43% in 2025, and cites Wood Mackenzie forecasting 10.8% average annual growth through 2034. Grid constraints and the search for dispatchable capacity, including the rush to build data centers, act as key drivers of this trend.
These numbers matter for CRE and charging providers because storage is increasingly being used to manage existing energy capacity through smoothing peaks, controlling ramp rates, and bridging short-term constraints while longer term capacity is secured.
Commercial sites look to microgrids
The data center playbook is also reviving microgrid architecture, not only for islanding during outages, but for day-to-day operational control. The section of the NERC assessment reporting on PJM notes an “extreme and rapid tightening” of capacity conditions and the need for additional resources to address near-term reliability challenges. This described the type of environment where flexible, controllable loads and behind-the-meter resources become more valuable.
Meanwhile, the legal and commercial tooling is catching up. Orrick’s data center guide explicitly points to co-located generation, behind-the-meter PPAs, and other contracting mechanisms used to secure power amid grid constraints and interconnection delays.
The overlooked lever: “grid-aware” operations
Software is making the load itself more flexible. A Dell Technologies / Forbes BrandVoice piece argues for making AI workloads more “grid-aware,” shifting non-urgent compute to times and places where grids are less constrained.
Even if most CRE doesn’t impact demand as much as computing power, it can still employ the same strategy. Dispatchability is becoming a design feature that includes managed EV charging, thermal storage, and building load controls.
How these trends affects CRE and EV charging
When data centers dominate load growth, the spillover is less about headlines and more about physics:
Shared infrastructure gets scarce. If a substation is being upgraded for a hyperscale customer, neighboring sites may face more complex studies, higher upgrade allocations, or longer queues. PJM is explicitly incorporating large-load increases (often data centers) into transmission planning due to resulting reinforcement needs.
Tariff and peak risk rises. As systems tighten, capacity and peak pricing pressures tend to increase—and “spiky” loads (like DC fast charging) look more expensive unless they’re controlled.
Behind-the-meter becomes competitive. The sites that can self-shape demand (storage and controls and flexible charging) will be easier to electrify and easier to finance.
The macro driver is already visible: Axios, citing an IEA analysis, reports data centers could account for roughly half of U.S. power demand growth for the rest of the decade, with overall U.S. electricity demand growth projected around 2% annually from 2026–2030.
How to adapt if you own or build on big electrical footprints
If you operate depots, campuses, industrial parks, or large retail/mixed-use with meaningful future load, plan the site with expected future competition for megawatts. Specifically:
Treat interconnection as a real estate risk factor rather than not a late-stage engineering task.
Design for controllability: managed charging, power-sharing, and site-level load caps.
Make storage financeable by stacking value: peak shaving, backup and tariff optimization and, where possible, grid services.
Leave room for a microgrid future: switchgear layouts, conduit pathways, and control system interoperability.
Data centers are just the first sector being forced into this playbook. Once the region’s biggest loads start treating solar and storage and microgrids as schedule insurance, everyone sharing the same wires eventually follows.
