Why power matters
See why electricity constrains compute.
← Back to Compute College
Compute College
What is a Data Center Interconnection Queue?
A data center interconnection queue is the waiting line for large facilities seeking electrical grid connection.
Plain-English definition
A data center interconnection queue is the waiting line and study process for large facilities that need to connect electrical load to the power grid. For an AI data center, queue progress can determine when a planned building and its GPUs become energized, operating compute capacity.
Memory trick: The interconnection queue is the checkout line before a data center can turn on: the equipment may be chosen, but it cannot leave powered and usable yet.
Why it matters
GPUs do not produce training or serving output without sufficient electricity. A developer may have land, financing, customers, and equipment plans while still waiting for utility studies, transmission upgrades, substations, permits, or physical connection. The queue can therefore set the real capacity schedule.
- Queue delays postpone usable supply even when market demand and chip orders are strong.
- Projects with faster power access may command greater value because buyers need capacity on specific dates.
- Grid bottlenecks connect AI compute expansion to regional infrastructure investment and policy.
Simple example
Imagine a company announces a planned 200-megawatt AI campus, while the utility indicates that full interconnection will not be available for three years. The 200 megawatts describe possible future scale; they do not represent near-term GPU supply for a buyer planning a job next quarter.
- Announced capacity and energized capacity must be labeled separately.
- Projects may phase power in over time rather than receiving the full stated load at once.
- Alternative power arrangements require their own reliability, cost, and timeline evaluation.
Example figures are illustrative calculations, not current quoted market prices.
Market signal
How to read the market signal
Longer queue timelines, required upgrades, delayed energization, or large-load constraints can signal tighter future AI supply. Faster approvals, completed upgrades, or credible alternative generation can suggest that planned capacity may enter the market sooner.
- Power milestones often matter more than construction announcements for near-term compute access.
- Regional queue pressure can influence where providers deploy new clusters and what buyers pay.
- A delay can preserve scarcity even when hardware manufacturing improves.
Market read: the relevant question is not only how much a campus plans to build, but when the site can turn on and serve buyers under dependable power conditions. Figures here are illustrative unless explicitly sourced and dated — see our methodology.
Common mistake
Do not treat a press release about facility scale as current supply. Planned capacity, under-construction capacity, partially energized capacity, and fully operating capacity are distinct market states. A project waiting for electricity is not yet available compute.
Practical takeaway
What you can do with this
Track power status alongside GPU orders and facility construction. Buyers should request the energization status and contracted availability of the capacity they intend to use; analysts should attach timelines and caveats to all capacity interpretations.
- Analysts: maintain separate fields for planned megawatts, connected megawatts, energized load, and operating cluster availability.
- Buyers: ask when reserved capacity is powered and deliverable under the contract, not just when a campus was announced.
- Investors: watch utility approvals, transmission or substation work, and phased energization milestones.
- Operators: communicate constraints clearly rather than presenting pending capacity as ready supply.
Decision check: before counting a project toward available compute, verify connection status, power amount, energization date, equipment status, and buyer availability.
Compute College
Turn the lesson into a number
Use the GPU-Hour Cost Calculator, AI Training Cost Calculator, or Model Serving Cost Calculator.
Compute College track
Power & Data Centers
Step 15 of 17: What is a data center interconnection queue