Congestion
What happens when transmission lines can't carry all the cheap power that wants to flow, forcing prices to differ across locations.
Transmission lines are like highways with a fixed number of lanes. When more cheap power wants to move from one region to another than the wires can carry, the grid operator has to intervene: it backs down inexpensive generators on the sending side and calls up more expensive ones on the receiving side. That workaround has a cost, and congestion is the name for it.
The visible symptom of congestion is a price split. Where cheap power is bottled up — say, wind in West Texas that can't all reach the cities — prices fall, sometimes below zero. Where demand sits on the far side of the bottleneck, prices rise. The gap between the two is the congestion cost of the constrained path.
Congestion isn't a malfunction; it's the market telling the truth about the physical grid. It signals where new transmission would be valuable, where a battery could earn money absorbing trapped energy, and where siting a new plant would be a mistake. For anyone trading or operating assets, knowing which constraints bind — and when — is most of the game.
For the technical reader
In the dispatch optimization, congestion arises when a transmission constraint binds — flow on a monitored element reaches its limit for a given contingency — and picks up a nonzero shadow price. Each bus's congestion component of LMP is the sum of binding-constraint shadow prices weighted by that bus's power transfer distribution factors (shift factors), so buses on opposite sides of a constraint move in opposite directions.
Congestion rent — the difference between what loads pay and generators receive across a constrained interface — is collected by the ISO and typically returned to holders of financial transmission instruments (CRRs in ERCOT, FTRs elsewhere), which exist precisely to hedge this locational spread.