How to Read a Grid Interconnect Queue

A developer signs a site, lines up capital, and files an interconnection request. The utility responds with a queue number. It is easy to read that number as a place in line, the way you would read a ticket at a deli counter. It is not. A queue position tells you almost nothing about when, or whether, the site gets energized. Reading it correctly is the difference between a project that delivers power in 2027 and one that quietly dies in a study cycle two years from now.

The Queue Is Not a Line

As of the most recent Lawrence Berkeley National Laboratory "Queued Up" analysis, roughly 2,600 gigawatts of generation and storage capacity sit in US interconnection queues. That is more than twice the entire installed capacity of the US grid. The number is not a measure of what is getting built. It is a measure of how many projects are waiting, and most of them never connect.

Of the projects that entered queues between 2000 and 2018, only about 14 to 19 percent reached commercial operation, according to LBNL. The rest withdrew. A queue, in other words, is mostly a list of projects that will not happen. Your position in it is a position relative to a population that is going to thin out dramatically, on a timeline you do not control.

What FERC Order 2023 Actually Changed

In July 2023, FERC issued Order 2023, which replaced the old first-come, first-served interconnection process with a first-ready, first-served cluster study model. The distinction is the whole game. Under the old rules, projects were studied one at a time in the order they filed, and a single speculative project at the front could hold up everything behind it for years.

Under cluster studies, the utility batches projects into a study group, analyzes their combined impact on the grid, and assigns network upgrade costs across the cluster. Projects that are not ready, that cannot post financial commitments or show site control, get dropped from the study rather than allowed to stall it. The intent was to clear out speculative interconnection requests and move real projects faster.

For a data center operator, this changes what a queue position means. Being early no longer protects you. Being ready does. A project that filed later but has site control, a deposit posted, and a credible load profile can clear a cluster study while an earlier, softer project gets washed out. The question is not "where am I in line." It is "am I in the study group that is actually going to be built, and can I survive the cost allocation when it lands."

The Cost Allocation Surprise

The part of the cluster process that catches developers off guard is network upgrade cost allocation. When a study group is analyzed together, the cost of the transmission upgrades needed to accommodate the whole cluster gets divided among the projects in it. If several large projects in your cluster withdraw after the study, the remaining projects can inherit a far larger share of the upgrade bill than they modeled.

This is why a queue position taken in isolation is close to meaningless. The relevant questions are who else is in your cluster, how financially committed they are, and what the network upgrade exposure looks like if half of them leave. A good interconnection position is not an early number. It is a defensible spot in a cluster of serious projects with manageable shared upgrade costs.

How to Actually Read a Position

Reading a queue position means reading several things at once. Below is the data that actually predicts timeline, and what each item tells you.

Why This Decides AI Data Center Timelines

For most of the last decade, the binding constraint on a data center project was capital and real estate. In 2026, for large AI load, it is power, and power upstream of the meter means transmission and interconnection. A site can have land, a willing utility, and committed capital, and still be three to five years from energization because the local grid cannot absorb 100 megawatts of new load without upgrades that have to clear a study cycle.

This is the mechanism behind a pattern that shows up across the market: projects in congested transmission regions stalling while projects in regions with available headroom move. The queue is where that difference becomes visible. A position in a region with transmission headroom and a short study cycle is worth far more than an earlier position in a saturated one.

None of this is a secret to the utilities or the large developers. It is a secret to operators who read a queue position as a place in line and budget their timeline accordingly. The buyers who clear the credit filter and the interconnection filter at the same time are a small group, and the queue is one of the two places that group gets sorted.

The Practical Read

If you are evaluating a data center site in 2026 and someone hands you a queue number, the number alone should not move you. Ask what study phase it is in, what the cluster around it looks like, what the network upgrade estimate is, and whether the local grid has the headroom to take the load without a multi-year transmission build. A late position in a clean, low-congestion cluster with a posted deposit beats an early position in a saturated market with an unbounded upgrade bill, every time.

Interconnection has moved from a back-office step to the front of the site decision. The projects that get built on schedule are the ones that treated the queue as a design input from the first site visit, not a formality to file after the deal closed.

Further reading

• The Credit Filter Nobody Explains - the other screen that decides who gets capacity in 2026.
• Where Midmarket Compute Actually Deploys - why secondary markets with grid headroom are absorbing demand.