DOE Emergency Grid Order Defies Court Ruling, Targets Duke Energy

The U.S. Department of Energy issued a rare emergency order under the Federal Power Act on 18 September 2026 to shore up the Carolinas grid, less than a week after a federal appeals court ruled that the DOE had exceeded its authority doing exactly the same thing in Michigan.

The timing frames the story. Order No. 202-26-48 took effect on 18 September and runs through 21 September, directing Duke Energy Carolinas to keep specific generation on standby. It landed alongside a parallel federal push to keep Indiana coal plants running, and just seven days after the D.C. Circuit vacated a structurally similar order for the J.H. Campbell coal plant. This is not a routine grid notice. It is a federal policy pattern colliding with judicial limits in real time.

What follows here matters because the DOE’s accelerating use of its emergency power speaks to two things at once: how thin the grid’s operating margin actually is right now, and how durable traditional generation assets are heading into fall. The near-term operational stakes and the longer-term policy signals both feed directly into how energy investors should read the current environment.

What the DOE’s Carolinas order actually requires Duke Energy to do

The mechanism here is narrow by design, and understanding how narrow it is tells you how tight the grid margin has become. Order No. 202-26-48 was issued under Section 202(c) of the Federal Power Act, the statute that lets the DOE compel generation during an emergency. It became effective on 18 September 2026 and expires on 21 September 2026.

The order authorises Duke Energy Carolinas and its Transmission Owners to dispatch specified generating units and activate designated backup generation. Crucially, this is not a blanket instruction to run plants. It is a last-resort provision, triggered only before or during an Energy Emergency Alert Level 3 (EEA-3), the highest tier of grid emergency, when a utility is at risk of shedding load involuntarily.

The DOE’s framing is deliberate: backup generation may be activated only as a “last resort” measure ahead of or during an EEA-3 event. That precise language carries legal weight, and it becomes important later.

The order followed a formal application Duke submitted to the DOE on the same day, 18 September 2026. What makes this significant is that it is the second such directive for this territory in September alone.

The three Carolinas-territory orders this year, in sequence:

• 11-12 June 2026: Authorised maximum output on specified units notwithstanding air-quality or permit limits, referencing anticipated load of approximately 34,589 MW.
• 3-8 September 2026 (Order No. 202-26-43): Addressed sudden demand increases and shortages, authorising temporary unit dispatch and backup generation.
• 18-21 September 2026 (Order No. 202-26-48): The current order, again permitting last-resort dispatch ahead of an EEA-3.

Two emergency orders for the same territory inside three weeks tells you something the DOE does not say directly: the Carolinas grid is running without enough dispatchable margin to absorb routine late-summer heat, not just catastrophic events. The DOE estimates more than 35 GW of untapped backup generation capacity nationally. For investors in peaking and backup capacity assets, repeat issuance like this is a demand signal worth registering.

Indiana’s coal plants and the federal pattern of keeping legacy generation alive

The Carolinas order is not an isolated regional response. It is the same statutory instrument the DOE is deploying elsewhere, on a different fuel type, for a different purpose, and that is where a coincidence becomes a strategy.

The PJM emergency orders issued earlier in September 2026 follow the same Section 202(c) instrument and reinforce the pattern: the DOE is deploying this statute simultaneously across multiple independent grid regions, not sequentially in response to isolated local events.

In the Midcontinent Independent System Operator (MISO) region, the DOE has issued a series of Section 202(c) orders keeping three Indiana coal units available, explicitly citing resource adequacy and projected capacity deficits. The assets are the R.M. Schahfer Generating Station operated by NIPSCO and the F.B. Culley Generating Station operated by CenterPoint Energy.

The timeline is the tell. The initial orders were issued on 23 December 2025, halting planned year-end retirements and mandating operations through 23 March 2026. From there, the DOE extended in successive 90-day increments: first through 21 June 2026, then through 19 September 2026, with Schahfer subsequently pushed out to 18 December 2026.

That is one full year of continuous operation for plants that were slated to close. What began as a short-term emergency measure has become something closer to a standing commitment.

The administration’s stated rationale

Secretary of Energy Chris Wright has attributed the grid’s vulnerability during extreme temperatures to prior administrations’ decisions to decommission traditional power sources ahead of schedule. That framing is the political logic driving repeat orders: the argument that firm capacity was retired faster than replacement could be built, and that keeping legacy plants online is a pragmatic bridge.

For investors, the read is direct. A one-year lifespan on a supposedly temporary intervention signals these coal assets are being treated as structural capacity backstops, not stopgaps. That reshapes the calculus for both coal operators, who gain revenue visibility, and the clean-energy projects competing against federally protected legacy plants.

The D.C. Circuit just ruled this tool is being overused, and that ruling is already in effect

Here is the complication the DOE is pressing through. On 11 September 2026, the D.C. Circuit Court of Appeals vacated a DOE emergency order that had kept the J.H. Campbell coal plant in Michigan open past its scheduled retirement.

The court’s grounds were specific. It found the DOE had exceeded its Section 202(c) authority and had failed to demonstrate an actual emergency, which the statute requires.

The court held that preventing market-driven retirements in order to advance coal preservation is an improper use of emergency powers. That is the controlling legal language, and it applies to the same statute the DOE invoked for the Carolinas seven days later.

The D.C. Circuit ruling on coal retirements handed down on 11 September 2026 applied specifically to the J.H. Campbell plant in Michigan, but its statutory language reaches every active Section 202(c) order the DOE has issued using the same justification.

The contrast is stark. A federal court ruled on 11 September that using 202(c) to block market-driven retirements is not a qualifying emergency, and the DOE issued fresh orders in Indiana and the Carolinas anyway. The gap between the ruling and the new directive was one week.

The escalation in usage is what makes this ruling consequential:

1. Pre-2000: Section 202(c) was used roughly 29 times, most tied to World War II.
2. 2000 to mid-2025: Used approximately 20 times, predominantly for acute weather events lasting one to four days, averaging around once per year.
3. The recent period: Dozens of orders in under two years, approximately 40 times the prior annual average. Some trackers cite 72 orders in a 19-month span, though that figure is unverified.

Critics have put a number on the cost. An Earthjustice-cited study estimates that broad extensions of stay-open orders for large fossil plants slated to retire by the end of 2028 could add more than $3 billion in utility customer costs.

For anyone holding coal or gas-peaking exposure, the legal status of these orders is now a direct risk variable. A successful challenge to any current order could accelerate retirements and strand mandated capacity in weeks, not years. Any thesis built on federally extended legacy generation is now contingent on litigation, not just grid conditions.

What heat, solar fade, and shrinking reserve margins mean for grid reliability heading into fall

Behind the legal and political drama is a physical grid under measurable strain, and understanding the mechanics explains why these orders keep multiplying. Late-summer heat compounds stress from two directions at once: it drives cooling demand up while cutting the output and efficiency of thermal plants.

According to NERC’s 2026 Summer Reliability Assessment, released on 19 May 2026, net internal demand rose roughly 10 GW year-over-year, from about 780 GW in summer 2025 to approximately 790 GW in summer 2026. The driver is electrification and data-centre load growth.

The grid capacity crisis driving these orders is not a seasonal anomaly: demand from data centres, EV charging, and industrial electrification has been compressing reserve margins across multiple regions since at least mid-2025, well before the late-summer heat events that triggered September’s directives.

The grid did add supply, and at record pace. More than 58 GW of new generation has come online since summer 2025, including 16.4 GW of solar and 14.7 GW of battery storage. But demand grew alongside it. The grid is changing composition without becoming more resilient in absolute terms, which is what makes late-summer stress a structural feature of this transition rather than a one-off.

The efficiency losses sharpen the picture. Grid experts note solar panels can lose up to 15% of their output in high temperatures, and modelling finds average reserve margins fall by roughly 22% on heat-wave days compared with normal summer days. A McKinsey analysis models an extreme-heat scenario slashing reserve capacity from 44% to 3%, as thermal plants shed about 1.6 GW and renewable output drops 5.1 GW, leaving just 2 GW of reserves. Both figures are modelled scenarios, not confirmed grid-specific outcomes.

Why evenings are the riskiest moment on the grid

The most acute strain does not arrive at midday. It arrives after sundown, when three factors compound:

• Solar generation fades as daylight shortens into fall.
• Ambient temperatures and air-conditioning demand stay elevated.
• Wind speeds frequently drop at the same time.

When solar output collapses but demand holds, the grid leans hard on remaining dispatchable thermal units and battery storage to cover the rising evening “net load.” That is the exact window an EEA-3 alert is designed to manage, and the exact reason last-resort dispatch authority matters. For investors, the takeaway is that demand growth from electrification and AI infrastructure is outpacing grid buildout, sustaining pricing power for dispatchable generation of any fuel type, at least in the near term.

What the DOE’s accelerating emergency orders tell investors about the grid’s near-term direction

Pull the threads together and the tension is unmistakable. The federal government is using emergency powers at an unprecedented rate to keep traditional generation online, the judiciary has just ruled the approach exceeds statutory authority in at least one case, and demand is growing faster than firm dispatchable capacity is being added.

For coal and gas peaking assets, that creates a split outlook. Federally mandated operation gives plant operators genuine revenue visibility in the short term. But the 11 September 2026 D.C. Circuit ruling means that visibility is contingent on litigation outcomes now moving against the administration.

An Earthjustice-cited study estimates broad stay-open extensions through 2028 could impose more than $3 billion in additional utility customer costs, quantifying the ratepayer risk that sits underneath the legal fight.

The more defensible long-term position sits elsewhere. Grid modernisation, interregional transmission, and battery storage are the asset classes that would reduce dependence on emergency orders if built at scale, and the DOE’s own estimate of more than 35 GW of untapped backup capacity nationally points to how much slack the system is currently improvising around.

Interregional transmission constraints are a structural reason why surplus generation in one grid region cannot relieve stress in another, a limitation that makes local emergency orders a repeated necessity rather than an exceptional intervention during heat events.

Three near-term variables are worth watching:

• The outcome of pending 202(c) litigation following the D.C. Circuit ruling.
• The Schahfer order expiry in December 2026, the nearest active order to lapse.
• NERC’s fall reliability outlook.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Financial projections and modelled scenarios are subject to market conditions and various risk factors, and past performance does not guarantee future results.

A stressed grid and a contested legal tool, where the Carolinas order fits in the bigger picture

Step back from the individual order and it becomes one data point in an accelerating pattern. Dozens of Section 202(c) orders have been issued in under two years, across multiple regions and asset types, at roughly 40 times the prior annual average of a statute used about 20 times between 2000 and mid-2025.

The dilemma underneath is real, not manufactured. Demand from electrification and AI infrastructure is climbing, net internal demand reached roughly 790 GW this summer, legacy firm capacity is retiring faster than expected, and record renewable additions do not fully substitute for dispatchable generation during peak stress windows.

As of 19 September 2026, three active order territories illustrate the reach:

• The Carolinas (Order No. 202-26-48).
• Indiana’s Schahfer units.
• Indiana’s Culley unit.

The question the courts will now answer

The 11 September 2026 ruling marks the inflection point. Whether the federal government can legally sustain emergency orders as a substitute for long-term grid planning is now a matter for the courts, not the DOE alone.

That is the defining tension of U.S. grid policy this fall: the legal authority to run this strategy is contracting at the same moment the operational pressure justifying it is growing. For investors, the outcome will shape the environment for generation, transmission, and storage assets alike over the next 12 to 24 months.