The World's Largest Data-Center Market Has No New Gas in the Queue

Everyone knows PJM's Dominion zone, Northern Virginia, holds the largest concentration of data centers on earth. Less understood is how little new gas is actually coming to serve it. On PJM's own load forecast, the Dominion zone leads all of PJM in summer peak growth, roughly +6.7 GW to 2030. On the supply side, the entire PJM interconnection queue holds only about 0.7 GW of gas that is signed, near-term, and still to be built.

The queue is not how PJM builds firm power

PJM is a capacity market, not an energy-only merchant market like ERCOT. A developer does not need a merchant queue position to earn on firm capacity; the capacity auction pays for it. So near-term firm power leans on the existing gas fleet (about 95 GW across PJM), the capacity auction, and behind-the-meter arrangements. New gas still interconnects, but in PJM the queue is a weaker near-term supply signal than it is in ERCOT, and it understates the firm power the fleet and the auction already provide.

That does not make the constraint less real. PJM capacity-auction prices have moved up sharply as data-center demand pulls on a fleet that is not growing fast, and Northern Virginia, the zone with the most demand, has almost no deliverable new gas of its own.

What this changes for the buyer

Hyperscaler energy leads. In PJM, the firm-power question is a capacity-market and existing-fleet question, not a queue question. Underwrite the capacity cost and the fleet's available headroom, and treat behind-the-meter generation as the faster path where the queue and the auction cannot deliver on your timeline. Northern Virginia buys you fiber, latency, and an existing ecosystem, at some of the highest and most volatile power costs of any major market.

Infra funds. A merchant new-gas thesis has little room in PJM, because the capacity market and the incumbents, not new queue entrants, meet the load. The investable firm-power angles are existing assets, uprates, and behind-the-meter structures, not a speculative new-build queue position.

Methodology

The 0.7 GW deliverable figure is reconciled from PJM's public New Services Queue (the PlanningQueues feed): gas projects that are progressing, hold an executed interconnection agreement, and target a 2026 to 2028 in-service date, excluding withdrawn and already-operating projects. It is Tafel Power's filtered estimate, not a PJM-published category. Demand growth by zone is from PJM's 2026 Load Forecast Report. The 95 GW existing-gas figure is from EIA-860M, grouped by balancing authority. Figures reflect mid-2026 snapshots and may have changed since.

All data compiled by Tafel Power from public sources. Framing informed by the firm's transaction advisory work in ERCOT and cross-ISO markets.