How the Queue Got Clogged

If the queue is the clock, the natural next question is why the clock runs so slowly. The queue did not congest because grid operators are lazy. It congested because a process designed for a slow trickle of large power plants was suddenly asked to absorb a flood of projects it was never built to handle, and its own rules made the flood worse.

Start with the volume. Over the last several years the number of projects requesting interconnection has grown enormously, driven by cheap solar, batteries, and now very large computing loads. In many regions the megawatts sitting in the queue now exceed the entire installed capacity of the grid they want to join. A process that once handled a handful of thick, serious applications per year began receiving hundreds, many of them thin and speculative.

That speculation is the second cause. For a long time it cost very little to enter the queue. A developer could file on a dozen sites, hold the positions cheaply, and wait to see which one drew the best study results before committing real money. The queue filled with options rather than commitments. Because operators still had to study each request as if it were real, engineering effort was spent on projects that were never going to be built — and every one of those phantom projects sat ahead of, and distorted the studies for, the projects that were real.

Now combine speculation with the interdependence I described earlier, and you get the mechanism that truly broke the system: the restudy cascade. Under the old serial approach, each study assumed every project ahead of it would be constructed. When a speculative project near the front withdrew — as many did once it saw its cost assignment — the removal changed the power flows for everything behind it, forcing restudies. Restudies took months, during which more projects withdrew, triggering more restudies. Operators found themselves running on a treadmill, redoing analysis faster than they could finish it. Queue times stretched from a couple of years toward five and beyond in the worst-hit regions.

Cost allocation made withdrawals worse. Because a late-arriving project could be assigned the full cost of a large upgrade that also benefited others, developers faced enormous, unpredictable bills that landed only at the facilities-study stage — after years of waiting. Rational developers walked away, and each walkaway restarted the cascade for their neighbors. The very structure that was meant to make projects pay their fair share instead manufactured instability.

The fix now being implemented across the country is a shift from serial, first-come study to "cluster" study. Instead of examining projects one at a time, the operator gathers a window of applications into a group, studies them together, and allocates the shared upgrade costs among the group by how much each contributes to the need. Speculative entry is discouraged with higher deposits, financial commitments, and readiness requirements — proof of site control, for instance — so that fewer phantom projects clog the line. Projects that drop out forfeit deposits and, crucially, the study is designed to be more robust to their departure, taming the cascade. This is the thrust of the federal reform to the generator interconnection process finalized as Order 2023: study in clusters, make entrants demonstrate they are serious, and stop letting one withdrawal reset everyone's clock.

For an investor, the transition itself is information. A region that has adopted a well-run cluster process and cleared its legacy backlog is a fundamentally different risk than one still grinding through serial restudies. Ask which regime a site sits under; the answer moves the timeline by years. For an operator, the reform changes tactics: readiness now buys you position. Because the process rewards demonstrated site control and financial commitment and punishes optionality, the developer who shows up with genuine readiness moves through a cluster faster than the one hedging across many speculative filings.

It is worth holding the empirical shape of this in mind. The queue clogged not from a single villain but from a feedback loop — low entry cost invites speculation, speculation inflates volume, volume plus interdependence produces restudies, restudies produce withdrawals, withdrawals produce more restudies. Reform works by attacking the loop at several points at once: raising the cost of frivolous entry, studying in groups so the analysis is more stable, and sharing upgrade costs so no single project faces a ruinous surprise bill. Understanding that loop is what lets you tell a queue that is genuinely improving from one that is merely promising to.