The United States
The seven ISOs, and the third of the country that never joined one. PJM, MISO, CAISO, NYISO, ISO-NE, SPP, with the regulated Southeast and West as a deliberate control group.
Words the industry uses are marked like this at the sentence that defines them, so you can tell a term you will hear on a desk from a phrase this book happens to be using.
There is no American electricity market. There are seven, covering around 70% of demand, and a large remaining territory that has no market at all. The boundaries between them follow the history of which utilities agreed to pool with which neighbours, which is why they cut across state lines, ignore geography, and align with the physical interconnections only loosely.
A single company in Pennsylvania sits inside a market spanning thirteen states. A company in Georgia, four hundred miles away, dispatches its own plants for its own customers and has no wholesale price at all. Nothing about the physics or the geography explains the difference. The restructuring wave reached one and stopped before the other.
Table 17-1: The seven, and what distinguishes each
| Operator (ISO/RTO) | Roughly | The defining feature |
|---|---|---|
| PJMPennsylvania, New Jersey and Maryland | Mid-Atlantic and Ohio Valley, 13 states and DC | Largest by demand. A capacity market that has become a political crisis |
| MISOMidcontinent Independent System Operator | Midwest down to the Gulf | Long and thin, so its constraint is moving power north to south |
| CAISOCalifornia Independent System Operator | Most of California | The duck curve in its purest form, and the hub of the western expansion |
| ERCOTElectric Reliability Council of Texas | About 90% of Texas | Its own interconnection, energy-only, outside federal jurisdiction. Chapter 18 |
| SPPSouthwest Power Pool | Great Plains, now into the West | Wind-dominated, and the first RTO spanning two interconnections |
| NYISONew York Independent System Operator | New York State | A constrained city at the end of the system, so location matters enormously |
| ISO-NEISO New England | Six New England states | Winter fuel security, because the nat gas pipelines into the region are full |
PJM is the clearest illustration of the point above. Those three states are the ones whose utilities had pooled together by 1956, and the market now covers thirteen states and the District of Columbia.
The split that causes most of the trouble
American electricity is governed by two authorities: federal and state.
Federal (national/country level) regulators have jurisdiction over wholesale sales and transmission in interstate commerce. That authority sits with the Federal Energy Regulatory Commission, FERC, which Chapter 2 described being created in 1977. FERC covers the markets themselves: how they clear, what the rules are, what a capacity auction looks like.
Individual states retain authority over retail rates, over where generation may be built and permitted, and over what their utilities must procure. That authority is exercised by a state regulator, called a public utility commission (PUC) in some states and a public service commission (PSC) in others. Both are usually called the commission by anyone dealing with one. A state can decide it wants offshore wind, or no new nat gas, or a nuclear plant kept open, and it can do so regardless of what the wholesale market would have selected.
The arrangement worked while demand growth was ultra low and prices relatively stable. It has become unstable now that federal market outcomes drive state retail bills upward while state policy drives what is available for the federal market to clear.
A third body appears throughout this book and belongs in a different category from either half of that split. The North American Electric Reliability Corporation, NERC, writes the mandatory standards the bulk power system is operated to. It is not a government agency and it does not regulate any market. It is a non-profit that FERC certified as the Electric Reliability Organization under the Energy Policy Act of 2005, passed after the 2003 cascade blackout across the American Northeast which began with sagging lines in Ohio contacting trees, described in Chapter 3 and not prevented by the voluntary arrangements then in place. FERC approves its standards and NERC enforces them through regional entities, with penalties the statute allows to reach a million dollars a day for each violation. Its territory is not the United States: it covers Canada and a corner of northern Mexico as well, so a Canadian utility follows standards written by a body an American regulator certified.
That distinction matters for the next chapter. ERCOT in Texas sits outside FERC's jurisdiction over wholesale markets, and NERC's reliability standards still apply to it through a regional entity called Texas RE.
PJM, and the argument about who the market belongs to
PJM is the largest of the seven and is where the tension due to mid 2020s load (demand) growth has arrived first.
Chapter 16 covered the capacity prices. Stated as annual totals rather than dollars per megawatt-day, PJM's capacity cost went from about 2.2 billion dollars in 2023 to roughly 14.7 billion in 2024 and about 16.1 billion in 2025. For scale, 16 billion dollars is roughly 250 dollars a year for every one of the 65 million people PJM serves, spread across households, shops and factories rather than falling on any one of them, and Chapter 16 puts the effect on a retail bill at somewhere between 1.5 and 5% depending on the state. Those are costs that flow through to retail bills in thirteen states, set by an auction run by PJM, an organisation none of those states control.
The political response has been unusual for crossing party lines completely. Governors of both parties have threatened to take their states out of PJM's markets unless states are given a formal role in its governance. Pennsylvania's governor put it plainly at a state-led conference, saying that if PJM refused to change, the states would be forced to go in a different direction. Maryland has gone further and legislated a study, requiring its Public Service Commission and state level energy administration to examine the costs and benefits of leaving PJM entirely, or of leaving only its capacity market, with a report due at the end of 2026.
The structural complaint underneath the politics is real. PJM's capacity auction determines a cost that state electorates pay, while the states decide what may be built and where. An institution accountable to its members is producing a number that voters hold their governors responsible for. That is a governance mismatch rather than a market failure, and no adjustment to the auction resolves it.
The data centre load in Chapter 25 is what turned a long-standing structural awkwardness into a crisis, by making the number large enough to notice. The power market is responding with additional supply, and now has the economic incentive to do so, but that same incentive over the short term becomes political.
The third that never joined
Across the Southeast, much of the Mountain West and parts of the Northwest, roughly 30% of the US market, vertically integrated utilities dispatch their own fleets without a wholesale market price, as Chapter 14 described. These utilities set tariffs that are approved on a state level basis. The question of whether they should form a market has been argued for twenty years, so compare the first steps recently taken in the US Southeast and the Mountain West.
Rather than an organised market, southeastern utilities built the Southeast Energy Exchange Market, an automated platform allowing bilateral trades in fifteen-minute intervals using transmission capacity that would otherwise go unused. It has around two dozen participants including Duke, Georgia Power and Dominion in South Carolina. It was approved by a divided federal commission, and that approval was later remanded by a federal appeals court.
The operating results for this Southeast market tell the tale. In 2023 the platform produced about 3.7 million dollars of benefits against roughly 4.3 million dollars of administrative cost.
Compare that to the figure from Chapter 14. The Western Energy Imbalance Market, which does something similar in concept, has delivered around 5.6 billion dollars of benefits since 2014. The difference between a few million and several billion is not explained by geography or by the technology available. It comes from how much the design actually lets the participants share. The Southeast platform permits trades only where the parties would have transacted anyway, over capacity nobody else wanted, and captures very little. The Mountain West version is a real-time market that centrally dispatches across balancing areas and is designed to interface with much larger economic values.
The queue as the actual constraint
For anyone trying to build in the United States, the constraint has stopped being the cost of generation and become the wait for permission to connect.
Total interconnection queue capacity across the country ran to roughly 2,600 gigawatts in 2023 and about 2,300 gigawatts in 2024, the first decline in at least a decade. This compares to national peak demand of something over 700 gigawatts, a queue several times the size of the current system.
Federal reform arrived in 2023, replacing the first-come first-served serial study process with a first-ready first-served cluster approach. Projects are now studied in groups rather than one after another. So the cost of the network upgrades a group of projects collectively requires, a new transmission line or a larger substation, is identified and divided among them at the study stage, rather than arriving years later on whichever project happened to trigger it and moving again every time one of them withdrew. Applicants must also demonstrate site control and financial readiness to hold a place. The 2024 decline in queue volume is the first evidence of speculative projects being filtered out.
None of this is a shortage of money. A transmission owner earns a regulated return on what it builds, as Chapter 14 sets out, and utility capital has moved toward wires for that reason. What is scarce is permission. A line crossing several states needs consent from every one of them, and a state that captures none of the benefit can decline to give it. Who should pay for a line whose benefits are spread across a region is genuinely unsettled, and FERC has now made two attempts at answering it, Order 1000 in 2011 and Order 1920 in 2024. Behind both sits the equipment queue of Chapter 26, which no approval shortens.
The New England Clean Energy Connect is the case to read. It is 145 miles of line built to carry Quebec hydro into Massachusetts, at a cost of about a billion dollars. Maine voters approved a referendum against it in November 2021, by 59%, and construction that had already begun stopped. The Maine supreme court found the referendum could not constitutionally be applied to a project already under way, a jury found in April 2023 that the developer had vested rights, and the line entered service in January 2026, about a decade after it was first proposed. Every one of those years was spent on permission rather than on engineering or on money.
Texas adopted a similar system. ERCOT hit the same wall from the demand side and answered it the same way, replacing individual processing of large loads with a batched study, as Chapter 18 describes.
What the map teaches
The boundaries in the US reflect which utilities pooled together decades ago. The market gaps reflect which states were watching California in 2001. Texas is separate because of a lawsuit over a high-voltage switch a utility closed on purpose in 1976 to push its grid into federal jurisdiction, which is where Chapter 18 begins. And the present pressure comes from a wholesale market architecture built in the 1990s, for a system whose demand had been flat for a generation, now being asked to allocate the cost of a load boom among state electorates who did not choose it and cannot vote on the institution producing the number.
Every one of the three axes from Part Three is being reopened in the United States at once. Who dispatches is being contested in the west and in the PJM governance fight. How location is priced is being settled inside the markets and absent outside them. How capacity is paid for is the subject of a capacity auction that has quadrupled in cost and prompted several states to study the exits.