On October 12, 2026, the U.S. Department of Energy (DOE) issued Emergency Order No. 202-26-50 under Section 202(c) of the Federal Power Act, directing PJM Interconnection to dispatch specified generating units to preserve grid reliability. The order is effective October 13 through October 16, 2026, and was issued in response to an application PJM submitted the day before, on October 11. For four days, the operator of the largest competitive wholesale electricity market in North America — serving 13 states and the District of Columbia — will operate under federal emergency authority.
That sentence deserves unpacking, because almost every word in it carries weight. An emergency order is not a press release. It is a legal instrument that temporarily rewrites the normal rules of grid operation: which plants run, which environmental limits apply, and who can be told to start a generator. This article explains what the order does, why DOE reached for it, and what a four-day mid-October window signals.
What the order directs PJM to do
The order has three operative pieces. First, it directs PJM to dispatch specified generating units and to order their operation as needed to maintain reliability. The units are those listed in Exhibit A to PJM’s application — DOE’s public summary does not name them, and that gap is worth stating plainly rather than filling with guesses. The mechanism is the point: normally PJM dispatches generation through its markets, and units run when their offers clear. Under the order, PJM can direct the listed units to operate when it judges their output necessary for reliability.
Second, the order authorizes PJM to direct backup generation resources to operate as a last resort — specifically, before declaring an Energy Emergency Alert (EEA) 3 or during an EEA 3. PJM’s own account makes the practical meaning concrete: it authorizes PJM to direct data centers and other large-load customers to switch to backup generation, if necessary. Backup generation means the on-site generators large customers already maintain — the diesel and gas units behind the meter at data centers, hospitals, and industrial sites. Directing them to run takes load off the grid at the moment of maximum stress. The “last resort” qualifier matters: this is a tool for the edge of an emergency, not a routine dispatch lever.
Third, the order authorizes PJM to call upon its Transmission Owners and Electric Distribution Companies to implement the order as needed. That is the unglamorous but essential plumbing: PJM operates the bulk system, while the wires companies own the facilities through which any dispatch instruction must travel. The order also provides temporary relief from environmental permit restrictions for generating units, according to PJM — a provision discussed in detail below.
What triggered it: a shoulder-season squeeze
The immediate cause was not a heat wave or a polar vortex. It was the calendar colliding with the weather. PJM is in outage season — the fall shoulder period, roughly mid-September through early December, when generator and transmission owners schedule required maintenance and construction because loads are lighter and the summer peak has passed. This year, that maintenance season ran into unusually warm fall weather that pushed demand well above typical October levels.
The numbers PJM published Monday afternoon tell the story. As of 4 p.m. Monday, PJM’s load forecast was 103,125 MW for Monday, 112,447 MW for Tuesday, and 112,282 MW for Wednesday. Against that demand, PJM expected roughly 62,000 MW of generation outages through the affected period, plus key transmission outages for maintenance and major construction. PJM had already recalled or canceled outages to the extent possible, including an October 11 recall of all active generation maintenance outage tickets effective noon on October 14. The 62,000 MW figure is what remained after those recalls — capacity that could not be brought back in time, or at all.
Around the order, PJM layered its standard emergency procedures: a Maximum Generation Alert and Load Management Alert for October 13, a Capacity Advisory extended through October 14, and a Maintenance Outage Recall for October 14. A Maximum Generation Alert tells transmission and generation owners to consider deferring maintenance so units and lines stay available, and notifies neighbors that exports from PJM may be curtailed. The Load Management Alert gives advance notice that demand response — enrolled customers paid in advance to reduce use during stress — may be called. PJM emphasized that the alerts require no action from ordinary customers. The Capacity Advisory deserves a note: it is a new Emergency Procedure endorsed by PJM members in June, created for approaching capacity shortfalls, and PJM used it for the first time during this episode.
What Section 202(c) actually is
Section 202(c) of the Federal Power Act is one of the oldest emergency powers in American energy law. Enacted in 1935, it authorizes the federal government to require, by order, “such temporary connections of facilities and such generation, delivery, interchange, or transmission of electric energy as in its judgment will best meet the emergency and serve the public interest.” The authority originally sat with the Federal Power Commission; when DOE was created in 1977, it moved to the Secretary of Energy. DOE may act on its own or on application, without prior notice or hearing. The trigger is broad: DOE may act during war, or whenever it determines an emergency exists due to a sudden demand increase, a shortage of energy or of generation or transmission facilities, a shortage of fuel or water, “or other causes.”
In 2015, Congress amended Section 202(c) to address the collision between emergency orders and environmental law. The amendment allows an emergency order to temporarily override federal, state, and local environmental requirements, while imposing conditions: an order that conflicts with environmental law is limited to 90-day increments, applies only during the hours necessary to meet the emergency, and must minimize adverse environmental impacts to the maximum extent practicable. Entities complying with such an order are shielded from liability for the resulting violations. This is the legal machinery behind the permit relief in the PJM order — and the reason emergency orders are controversial even when everyone agrees the lights must stay on.
A tool once rare, now routine
For most of its history, Section 202(c) was used sparingly. According to a Congressional Research Service report, DOE invoked the authority in response to 20 events between 2000 and June 2025, 11 of them weather-related. That pattern has broken decisively: since May 2025, DOE has issued more than 40 emergency orders and extensions under Section 202(c). The recent orders fall into two categories: retirement deferrals, compelling utilities to keep units online past planned shutdown dates; and active emergency dispatches, authorizing generators to exceed normal operating or environmental permit limits during acute stress. The PJM order is of the second kind — a dispatch order for a live squeeze, not a stay of execution for a retiring plant.
The expansion has drawn legal pushback. The most consequential challenge concerned the J.H. Campbell coal plant in Michigan, where DOE ordered the plant kept available past its planned retirement; in 2026, the D.C. Circuit rejected that use of Section 202(c) to block a retirement. That ruling does not touch the PJM order directly, but it marks the boundary: the authority is strongest when an operator asks for it during a real emergency, and weakest when used to override retirements no operator requested. This fall alone has seen repeated PJM orders, including an August order letting PJM run Unit 4 at the Wagner Generating Station in Maryland beyond its operating limit through November 17, and a September order (No. 202-26-45) built around another list of specified PJM resources. Order No. 202-26-50 is the latest entry in a season of emergency orders.
The EEA ladder — and why “before EEA 3” matters
To understand the backup-generation provision, it helps to know the alert ladder it references. Under NERC’s emergency operations framework, Energy Emergency Alerts communicate escalating stress among balancing authorities and reliability coordinators, and to the public. EEA 1 signals tight supplies and a potential resource deficiency. EEA 2 signals that the operator is using all available resources and that load management is imminent or underway. EEA 3, the highest level, means firm load interruption — controlled outages of customers who cannot be spared — is imminent or in progress. Demand response, voltage reduction, and emergency imports all come before EEA 3; firm load shedding is what EEA 3 announces.
By authorizing backup generation “as a last resort before declaring an EEA 3 or during an EEA 3,” the order places behind-the-meter generation in the final gap between voluntary and involuntary load reduction. The logic is straightforward: a data center running on its own generators for a few hours is preferable to a neighborhood losing power, and preferable to waiting until shedding is unavoidable to discover whether the backup fleet can be mobilized. The provision is, in effect, an instruction to use every megawatt of already-built generation before resorting to outages.
The environmental relief, and its limits
The environmental dimension is where the strongest objections arise. When DOE authorizes units to run beyond permit limits, the consequence can be emissions above permitted levels — sulfur dioxide, nitrogen oxides, and other pollutants. The statute’s answer is procedural: the order must be limited to the hours necessary to meet the emergency, must minimize adverse impacts to the maximum extent practicable, and expires after 90 days unless renewed. The PJM order’s four-day window is far inside that ceiling. Critics argue that repeated orders for the same region amount to a rolling waiver rather than a genuine emergency response. Defenders argue the alternative is worse: a permit limit honored during a shortfall means load shedding. The structural point is not in dispute: each 202(c) order temporarily transfers authority over environmental compliance from the permitting system to the operator’s real-time judgment.
What a four-day October window signals
The most important fact about Order No. 202-26-50 may be its date. October is not supposed to be emergency season. The emergencies that built the 202(c) canon were winter storms and summer heat domes — the peaks that define the planning year. A four-day order in mid-October says something different: the margin between available supply and demand has thinned to the point where an ordinary shoulder-season week can produce emergency conditions. PJM has said as much: its capacity emergencies historically occurred during systemwide Hot or Cold Weather Alerts, but it is now experiencing capacity issues during milder weather, and expects these instances to increase as demand growth outpaces new supply. The October order is that thesis in operational form — proof the system now has tight hours outside the peak season.
This is where the order connects to the longer adequacy story. PJM’s recent capacity auctions have cleared at FERC-approved price caps — $329.17 per MW-day for the 2026/27 delivery year, $333.44 for 2027/28, and $325 for 2028/29 — a market signal that accredited supply is scarce relative to the requirement. (Those are wholesale capacity prices, not retail bill impacts.) DOE’s own 2025 resource-adequacy analysis projected total electricity demand growing by roughly 101 gigawatts by 2030, including about 50 gigawatts from data centers. The October order shows how a tightening system announces itself: not with a single dramatic failure, but with a rising frequency of weeks in which the operator must reach for tools designed for emergencies.
Planning vs. operations: where this fits
Readers of this desk’s earlier explainers will recognize the distinction. Resource adequacy — the LOLE standard, reserve margins, capacity accreditation — is a planning discipline: it asks whether the fleet, on paper, can meet demand across thousands of simulated futures. A 202(c) order is an operations event: it says that in this particular week, with this particular set of outages and this particular weather, the planned fleet is not enough. A system can satisfy the 0.1 days-per-year LOLE standard and still need emergency orders in a bad week, because LOLE is an average over many simulated years, not a guarantee for any given week. A system that needs frequent emergency orders is telling its planners something: the simulations may be understating how often maintenance outages overlap with unseasonable weather. The October order is a single data point. The pattern of orders across 2025 and 2026 is the trend.
What to watch this week
Four questions will determine what this order means in retrospect. First, whether PJM actually declares an EEA, and at what level — an order that expires unused is a different story from one that precedes an EEA 3. Second, whether backup generation is actually directed to run, and which customers are called upon — which will shape the debate over who bears the burden of grid stress. Third, whether the order is extended or followed by further orders — rolling PJM orders through the fall would confirm this is a season, not a week. Fourth, what DOE’s required reporting reveals about how many hours the specified units actually ran. For now, the order is a fact, not a forecast. The grid is not failing. The story of this week will be written in how much of that authority proves necessary.