This week, David Mills, CEO of PJM Interconnection, published a letter to stakeholders that contained a sentence no grid operator wants to write: "The current situation is not tenable." 

PJM is not a fringe regional operator. It is the largest electric grid in the United States, serving 67 million people across 13 states and Washington D.C. The situation Mills is describing is a grid that can no longer keep pace with the explosive growth in electricity demand driven by AI infrastructure, data centers, and electrification, while simultaneously managing an interconnection queue that has collapsed under its own backlog, aging market design that was built for a different era, and rising prices that are straining both utilities and the customers they serve. 

This is not a regional problem. It is a leading indicator of what every major U.S. grid is working through. And for commercial and industrial facilities that have built their operations around the assumption of reliable, affordable grid power, it is a signal worth taking seriously. 

What Is Actually Happening at PJM 

The immediate trigger for PJM's white paper is a collision between two forces that have been building for years. 

The first is demand. Cloud computing and AI infrastructure have reversed decades of flat electricity consumption. Data centers in particular draw massive, continuous loads that do not fluctuate the way residential demand does. Demand in PJM's territory has grown to the point where the grid operator is warning of power shortages as soon as next year, and one of its largest member utilities, American Electric Power, which serves millions of customers across 11 states, is considering leaving the organization entirely. AEP's CEO was blunt on a recent earnings call: "The current state of PJM's performance does not give me great confidence that these issues will be resolved anytime soon." 

The second force is the supply side's inability to keep pace. PJM paused new interconnection applications in 2022, citing a backlog of years. In the period before the pause, more than 300 gigawatts of projects were in the queue. Only 23 gigawatts actually connected. Most developers withdrew rather than wait. Since the queue reopened, more than 800 new interconnection requests have been filed for 220 gigawatts of new capacity. The queue is longer than ever. 

PJM's white paper proposes three structural options, none of which are clean solutions. One would require longer-term supply commitments from utilities and generators, difficult in a market where solar and batteries can be installed in months, but natural gas turbines face multi-year backlogs. Another would effectively create tiered reliability, where customers who pay less could have their power cut first, a politically toxic option in an environment where utility bills are already rising, and legislators are threatening price caps. The third attempts to move PJM closer to real-time market pricing while preserving some long-term stability, a compromise that has already failed to satisfy the market's largest participants. 

TechCrunch's assessment of the situation is worth quoting directly: "It's looking like a messy few years ahead." 

Why This Matters Beyond PJM's Territory 

PJM's crisis is not unique. It is a leading indicator. 

The same forces driving PJM's problems, accelerating AI-driven demand, slow interconnection processes, aging infrastructure, and a market design built for a different era, are present in every major U.S. grid. ERCOT in Texas has warned that peak demand could more than triple by 2032. The Federal Reserve Bank of Dallas has projected a 50 percent increase in wholesale power prices within five years. At least 11 states are considering temporary bans on new data center construction because their grids cannot absorb the load. 

The pattern is consistent. Grid operators that worked well when supply exceeded demand are now operating in a fundamentally different environment, and the institutional and regulatory processes for adapting to that environment are not moving fast enough. 

For commercial and industrial facilities, the practical implication is clear. If the largest grid in the country is warning of power shortages next year, and its most significant member utility is evaluating whether to leave, the assumption that grid reliability will simply be there when you need it is no longer supportable. Grid dependence is a strategic risk, not just an operational inconvenience. 

The Three Things PJM's Crisis Tells Facility Operators 

First, the interconnection queue is not getting shorter. With 800 new requests filed and a backlog of years already in place, any facility planning to connect new generation or large new loads to the grid through PJM's interconnection process is looking at a multi-year timeline in the best case. The NEC 2026 changes to Power Control System requirements offer some relief by allowing facilities to right-size their electrical service based on managed demand rather than theoretical peak load, but they do not solve the fundamental queue problem. On-site generation and storage that avoids or reduces the interconnection requirement is increasingly the faster and cheaper path. 

Second, rising prices are structural, not cyclical. The dynamic PJM describes, unprecedented demand colliding with constrained supply and aging market design, is not something that resolves in a rate cycle or two. The Federal Reserve Bank of Dallas's 50 percent price increase projection reflects a structural supply-demand imbalance that will persist regardless of how PJM reorganizes itself. Facilities that lock in energy costs now through Power Purchase Agreements or on-site generation capture protection against that escalation that facilities relying entirely on utility rates will not have. 

Third, reliability can no longer be assumed. PJM's white paper explicitly raises the possibility of tiered reliability; some customers getting power cut before others in shortage conditions. Whether or not that specific proposal advances, it reflects a real underlying condition: grid reserve margins in PJM's territory have dropped to levels where prioritization decisions become necessary during stress events. A facility without islanding capability or backup generation has no protection if it ends up on the wrong side of that prioritization. 

What On-Site Energy Infrastructure Looks Like in This Context 

The facilities that are best positioned as the grid becomes less reliable and more expensive are the ones that have reduced their dependence on it, not eliminated it, but reduced it to the point where grid fluctuations do not determine operational outcomes. 

On-site solar generation displaces the most expensive grid power during daylight hours. Battery storage with intelligent dispatch reduces peak demand charges, provides backup during short grid events, and can be configured to maintain critical operations during extended outages through islanding capability. Together, they reduce a facility's grid exposure to a manageable baseline rather than a total dependency. 

For facilities in PJM territory specifically, the business case has become more urgent this week. AEP's potential departure from PJM would affect millions of customers across Ohio, Indiana, Michigan, West Virginia, Kentucky, Virginia, Tennessee, and Arkansas. The uncertainty that creates, in rates, in reliability, in long-term planning, is exactly the kind of environment where on-site energy infrastructure moves from a smart investment to a necessary one. 

The same logic applies across ERCOT, CAISO, MISO, and every other grid facing the same underlying pressures. PJM is simply the one that has been forced to say out loud what operators across the country are already experiencing. 

NextNRG's View 

The PJM white paper confirms something NextNRG has been building around for years: the assumption of abundant, reliable, and affordable grid power is no longer valid for facilities with critical operational requirements. 

NextNRG designs AI-driven microgrid systems built around on-site energy generation, battery storage, and intelligent energy management that reduces grid dependence, eliminates peak demand charge exposure, and maintains operational continuity during grid events. Our AI forecasting engine gives operators the visibility to prepare for grid stress conditions rather than react to them. Our Microgrid Controller provides islanding capability that maintains continuous power to critical systems within milliseconds of a grid event. 

The grid is not getting more reliable on a timeline that matches the urgency of most facilities' operational and financial planning. Building energy independence into your infrastructure now is the practical response. 

Contact the NextNRG team at nextnrg.com to discuss what on-site energy infrastructure looks like for your facility. 

Sources: PJM Interconnection white paper, May 2026; TechCrunch reporting, May 8, 2026; Federal Reserve Bank of Dallas, April 2026. This post is for informational purposes only and does not constitute investment advice. NextNRG, Inc. (NASDAQ: NXXT).