Most operators I talk to already know the number. Three to five years on a new interconnect, longer in some PJM and MISO zones, and the queue is not shortening. ERCOT has its own version of the same problem. Even the favorable utility territories have shifted from "we can probably help you" to "we have a study window opening in 2028."
That gap is a real business problem, not a slide-deck problem.
What the queue actually costs
The cost is not the wait itself. The cost is the deals that close around the wait.
A 30 MW request submitted in 2026 with a 2030 energization gives a competitor with even 5 MW of immediate capacity a four-year head start on the same AI training contracts. That is enough time for a hyperscaler to source elsewhere, sign a multi-year commit, and never come back to the queue applicant for anything time-sensitive again.
Power is roughly 40 percent of data center operating expense. The cost stack on top of power, cooling and staff and debt service, is meaningless if the power itself is on a four-year clock.
So operators are doing what operators do when the front door is locked. They are looking at the side doors.
What "off-grid" means in 2026
The term gets used loosely. Three things tend to sit under the same label, and they are not equivalent:
A backup generator farm. Diesel, propane, or natural gas units sized for outage continuity. Not the same as base load. Fine for tier-4 redundancy, not the answer for primary power.
A behind-the-meter generation project. Solar plus storage, sometimes with a gas peaker. Useful in certain geographies, capacity-limited in most.
A connection to an operating natural gas field with dedicated allocation. Wells, pipeline, separator units, and a gen-set sized to the campus. This one is rare because it requires a counterparty that owns the gas, not just a marketer who sells it.
The third option is what closes the queue gap, and it only works if the gas counterparty has actual operating infrastructure, not a development plan.
What a working version looks like
There is an operating field in Eastern Kentucky we have been pointing operators toward. The facts are public enough to share:
140 active wells producing today.
23 miles of pipeline already in the ground.
About 1 megawatt of continuous output running on the line right now.
Room to scale to 1 gigawatt plus on the same field.
In February of this year the field produced roughly 2,200 megawatt-hours equivalent. That is an operating number, not a projection. Land adjacent to the field is available for buildout. The regulatory environment is favorable, which in Kentucky energy terms means the state wants the project and is not creating obstacles.
The thing that makes this useful for an operator is the directness of the allocation. Gas is committed to specific projects at a set price per project. It is not a futures contract, not a hedge, not a unit of a portfolio. It is a physical commitment to a campus.
What it does not solve
Three things to be clear about, because we tell operators these in the first call:
It does not solve interconnect with the grid. If a campus needs grid backup or grid sell-back, that is a separate workstream and a separate timeline.
It does not solve permitting. State, county, and air permits still apply to combustion equipment, and those calendars belong to the state, not the gas owner.
It does not solve build. The operator still has to build, or hire someone to build, the actual plant. Power allocation is upstream of EPC, not a substitute for it.
What it does solve is the wait. Power that does not yet exist cannot be allocated. Power that is already flowing can.
What the math actually looks like
The frame that has been useful in CFO conversations is the present-value math on the queue delay itself.
If an operator's revenue model says a 20 MW campus generates X dollars per megawatt-month at full utilization, every month of queue delay is a discrete revenue gap. Multiply by the queue length. The result is what off-grid allocation has to beat on a total-cost basis to be worth pursuing.
It usually does, by a meaningful margin, once you include the cost of capital sitting idle waiting for an interconnect study to clear.
The math fails in two cases. First, when the deal economics depend on grid sell-back during low-demand hours. Second, when the campus already has interconnect approved and energization is less than 12 months out. Outside those two cases, the math runs.
The honest framing
I am not in the energy development business. Stone Path facilitates the introduction between data center operators and KYTX, the company that owns the field. The land next to it is owned by a separate landowner, also reachable through the same introduction. Gas is allocated project by project at the field owner's price.
The reason this is worth a 15-minute conversation rather than another whitepaper download is that allocation is finite. The field is operating, the capacity is real, and the first projects to get scoped will get scoped first. That is the whole urgency case. It is not a fake scarcity frame.
If a campus is two to four years from energization and the grid queue is the bottleneck, the math is at least worth running. If the queue is not the bottleneck, this is not the right tool.
The intro takes one email. The math takes about 15 minutes to scope.
Reach out through the off-grid power page and we will walk through whether the field's current allocation window lines up with the campus timeline.