Data center developers are actively exploring how to close the power gap that grid interconnection queues have created. With timelines stretching five years or more, behind-the-meter natural gas generation has emerged as one of the most commercially viable near-term options for AI-scale facilities. But committing to BTM gas doesn’t simplify your energy strategy. It trades one infrastructure constraint for another: pipeline access, basis exposure, firm transport capacity, and upstream supply reliability all become site selection variables on the same level as land cost and fiber connectivity.
This article is for data center developers working through that reality. Not the investment case for gas infrastructure, but the operational question of how to evaluate sites with supply confidence, structure gas contracts before construction commits you to a delivery point, and avoid the mistakes that are already showing up as BTM projects move from announcement to groundbreaking.
The grid interconnection problem is getting worse, not better
According to Enverus Intelligence® Research (EIR), queue-to-commercial-operation timelines have grown roughly 60% since 2017, now averaging over 2,100 days for projects with a first power year in 2025. Getting a new data center connected to the grid now routinely takes five or more years in constrained markets. Only about 10% of capacity sitting in interconnection queues will actually get built, per EIR analysis. That timeline is commercially untenable for AI workloads where compute demand is already live and capital is already committed.
The math pushes developers toward a different model: generate power on-site, consume it directly, and operate independently of grid availability. That’s the definition of behind-the-meter generation. Power is produced at or adjacent to the facility, behind the utility revenue meter, without routing through the traditional grid interconnection process.
What “behind-the-meter” actually means for data centers
Behind-the-meter (BTM) power in a data center context means the generation asset sits on-site or co-located with the facility. Power goes straight from turbine or engine to the facility load, never touching the utility grid. The data center operator controls their own power supply instead of depending on a utility queue or a grid interconnection approval.
This is fundamentally different from a standard power purchase agreement or utility tariff. With BTM generation, the operator owns or contracts the generation asset directly. Speed to power improves significantly. Energy cost certainty improves. Reliability goes up.
The trade-off is capital intensity and operational complexity. Someone has to fuel, maintain, and run the generation fleet. That’s where natural gas supply strategy becomes a core part of data center operations, not just an infrastructure footnote.
Webinar: Data Centers for Operators: Gas Supply & Site Insights
Why natural gas is an attractive BTM fuel
Renewable BTM generation is theoretically appealing. In practice, it doesn’t work for most data center applications. A wave of hyperscalers are also exploring nuclear energy to secure power for data centers.
Solar and wind require large physical footprints that most data center sites can’t accommodate. Battery energy storage addresses short-duration backup needs, not continuous baseload. And combining renewables with storage to meet the 24/7, high-density power demands of AI compute at scale remains cost-prohibitive for most developers today.
Natural gas turbines and reciprocating engines solve the problem that renewables can’t: they deliver continuous, dispatchable baseload power at the scale data centers require, with deployment timelines short enough to matter.
What this means for gas supply strategy
Behind-the-meter power shifts data center operators from energy consumers to energy producers. That shift carries a set of gas supply responsibilities that the sector is still learning to manage.
A 100 MW data center running on natural gas generation might require several hundred million cubic feet of gas per year, depending on efficiency and load factor. To put it in basin terms: EIR estimates that a 1 GW data center consumes approximately 140 million cubic feet per day (MMcf/d) of natural gas, less than 1% of Appalachia’s daily production. Securing that volume at the right delivery point, with appropriate contract structure and price risk management, requires the same rigor that midstream operators and industrial gas users apply to their supply portfolios.
Key considerations include:
Basis risk. Data centers located near producing basins like the Permian, Haynesville, or Appalachia may face significant basis differentials versus Henry Hub. Understanding local price dynamics and securing supply at relevant delivery points is critical to controlling fuel cost.
Supply reliability. BTM generation doesn’t provide operational resilience if the gas supply is unreliable. Pipeline access, firm transport, and backup supply options need to be evaluated as part of site selection, not after construction begins.
Volume scalability. Data center power demand can grow rapidly as compute capacity expands. Gas supply agreements need to accommodate volume growth without punitive renegotiation terms.
Emissions accounting. Natural gas is increasingly framed by hyperscalers as a bridge fuel: temporary generation until grid power is available or until lower-carbon baseload alternatives reach commercial scale. How emissions from BTM gas generation are reported, offset, or managed under corporate sustainability commitments is a live question with no settled answer across the industry.
Site selection can be done along with available gas supply
The practical implication of the BTM shift is that natural gas access has become a primary site selection variable for data center development. In some markets, it now outweighs land cost or fiber connectivity.
Developers evaluating sites now need answers to questions that their energy teams weren’t asking three years ago:
- What gas gathering and pipeline infrastructure exists within economic connection distance?
- What are the firm transport options to the most likely delivery points?
- What is the basis differential history at this delivery point, and what price exposure does that create at scale?
- What gas producers or midstream operators are active in this basin with the volume and contract flexibility to serve a large BTM load?
- How does local gas supply interact with regional power market dynamics if the project eventually transitions to grid-connected operation?
These aren’t questions that real estate teams or network infrastructure teams can answer. They require deep familiarity with upstream production data, midstream asset maps, basis market history, and gas supply contracting. That kind of intelligence has historically lived on the energy side of the industry, not the technology side.
Natural Gas Transmission Analytics
For developers building behind-the-meter gas generation, supply confidence depends on what’s happening upstream of the meter. Enverus Natural Gas Transmission Analytics in PRISM® tracks daily flows across roughly 30,000 transmission meters on U.S. interstate systems, so you can trace a molecule from basin to the delivery point serving your site. See how full the serving pipeline is, where capacity is tightening, and how competing pulls from LNG, industrial load and other data centers could affect availability and basis pricing. Paired with the upstream production, midstream asset and power grid data already in PRISM, site selection, fuel-supply contracting and long-term cost exposure can be evaluated in one workflow rather than stitched together across vendors.
How Enverus helps
With Enverus, data center teams can:
- Find substations with real available power using queue, ATC and load interconnection data.
- Map competing data center projects to avoid contested interconnection capacity.
- Assess gas pipeline access for co-located or backup generation.
- Combine parcel-level land data with grid and gas infrastructure in one view.
