The 'Associated' Power Play: Monetizing Undervalued Natural Gas Resources to Fuel Hyperscale Data Centers

May 29, 2025
AI data center developers are turning to behind-the-meter solutions to sidestep grid bottlenecks and meet growing energy demand. Natural gas-rich regions like the Permian Basin are being leveraged for on-site generation, creating new revenue streams.

The hunger of artificial intelligence (AI) models for electricity is outpacing what utilities can timely deliver, threatening to overwhelm the grid and potentially exceed reliability standards.

Data center developers are getting creative and turning to behind-the-meter (BTM) solutions to sidestep grid bottlenecks. This pivot isn't just about avoiding transmission delays—it's redefining the industry around "speed-to-power.”

These developers are increasingly turning toward natural gas-rich regions like the Permian Basin, where abundant fuel supplies can power on-site generation right at the facility doorstep. The strategy creates new revenue streams for undervalued energy assets, while meeting tech's growing demand for energy—if developers can navigate the complex legal and operational landscape.

Stranded Gas and Renewables for Behind-the-Meter Projects

There’s an “associated” revenue opportunity emerging for producers of undervalued natural gas.  That opportunity comes from capturing natural gas that is in some instances burned or flared at the wellsite or otherwise sold downstream at steeply discounted prices.

The advent of combining hydraulic fracturing and horizontal drilling unleashed the shale revolution, leading to record production of crude oil and natural gas in the United States. Natural gas produced in conjunction with crude oil, otherwise known as associated gas, will be produced irrespective of its market price due to the value of crude oil.

In years past, associated gas was uneconomical to transport due to geographic isolation or lack of pipeline infrastructure. Producers burned, or flared, associated gas at the wellsite. Eventually, gas gathering, processing, and transportation systems were built to capture and move this associated gas to downstream markets for sale. 

Associated gas soon filled available pipeline takeaway capacity to higher priced markets, leading to depressed regional prices that at times reached negative pricing.  Likewise, in other parts of the country, natural gas produced on its own filled available takeaway capacity of several gas pipelines, resulting in depressed regional prices for gas that cannot be transported to more profitable markets.  

This is where the opportunity sits. BTM projects create an opportunity to convert natural gas that is “stranded” in oversupplied markets into electricity by using it as fuel for on-site natural-gas fired generators. These generators can directly power co-located data centers, bypassing grid interconnection. For example, a new collaboration in Appalachian Region will use natural gas and captured waste methane to power fuel cells generating up to 360 MW of electricity. 

There’s another revenue upside for producers: Natural gas sold to nearby to power producers may be able to bring a higher price than downstream purchasers and, avoid costly transportation costs.

Though promising, BTM power using associated gas or nearby produced natural gas alone cannot meet the projected 945 TWh global demand for data center energy by 2030. A more realistic strategy involves integrating gas-fired generation with other energy sources—like renewables and battery energy storage—along with connections to the grid as they become available. Pairing data centers with multiple generation sources allows data center operators to diversify their power supply and ease grid demand. Google and Intersect Power are applying this approach through a partnership to co-locate data centers with solar and battery storage, aiming to bridge the energy gap without relying on exclusively on transmission infrastructure.

Re-Imaging the PPA for the Data Center    

We’re entering a new age of power purchase agreements. To make this work, both the energy provider and the data center operator need to understand what goes into developing successful offtake agreements. These agreements typically include multi-year power purchase commitments, performance guarantees, and bankruptcy protections, such as letters of credit or other credit support and step-in rights. These features ensure revenue certainty for power producers and stable energy supply for data centers.

Power Purchase Agreements (PPAs) are central to this structure. Physical PPAs involve direct purchase and delivery of electricity from a generator to an electricity consumer at a fixed or indexed price.

Traditionally, the grid acts as the intermediary delivering the power from the generator to the purchaser, but BTM physical PPAs remove this intermediary and allow direct connection and delivery of power from the generator to the purchaser.

Virtual PPAs, by contrast, are financial agreements where the buyer pays a fixed price and settles with the market without taking physical delivery of the energy produced by the generator with the physical power being sold into the market or to a third party. By nature, virtual PPAs can only function with a grid connection. While more flexible geographically, virtual PPAs expose parties to market volatility, do not provide direct power to the purchaser, and rely upon interconnections to the grid that may require years to achieve. 

A data center, co-located with a power producer receiving BTM electricity from either a dispatchable thermal resource or an intermittent renewable resource that is connected or will eventually be connected to the grid will need a PPA with the attributes of both a physical PPA receiving BTM electricity as well as virtual PPA being financially settled in order to ensure that the power producer is capable of selling the entire output of its facility under a variety of conditions both behind the meter and on the grid.

Co-location agreements enable data centers to be built adjacent to power generation facilities, providing a firm power source and reducing reliance on the grid. For example, Diamondback Energy is actively seeking partnerships with data center operators to co-locate facilities in the Permian Basin, enabling the delivery of BTM power from natural gas fired generation. These arrangements can also reduce grid congestion costs by directly matching generation with consumption. However, regulatory uncertainty still clouds these arrangements as FERC looks to ensure grid reliability and equitable consumer pricing as more BTM projects are built while keeping in mind that the arrangements present many state regulatory issues outside of its jurisdiction.

Avoiding Transmission Congestion and Interconnection Delays

Grid integration is one of the biggest challenges for AI data center developers, particularly in high-demand markets like Texas. Congestion occurs when demand exceeds the capacity of a transmission line.  Left unchecked, a congested line may lead to reliability issues as the line becomes overused.  While regional transmission organizations (RTOs) or independent system operators (ISOs) like ERCOT physically manages the grid to avoid congestion, they also create markets to allocate congestion costs through financial mechanisms.  These congestion costs can be a significant portion of the total charges for electricity that an industrial customer pays each month. These costs are compounded by lengthy generator interconnection delays—despite ERCOT and the utilities within it having one of the nation’s fastest generator interconnection processes—it can still take years to navigate backlogs and permitting. Load interconnection processes, including data center load interconnection, is typically not subject to formal processes and without set processes and with utilities backlogged with requests, can create uncertainty and delay.  These delays create uncertainty for data center developers whose projects are often on tight deployment schedules to support rapidly expanding AI workloads and to control projected development costs.

By operating independently of an interconnection queue in an ISO or RTO like ERCOT, BTM systems (which do not need a generator interconnection) can be deployed much more rapidly, with potentially fewer regulatory hurdles and less exposure to grid congestion charges and potential instability, while at the same time reducing cost volatility. This allows data centers and other large industrial loads to move forward with development timelines and secure reliable power without waiting years for transmission buildouts. In effect, BTM power transforms what would be long-term development risks into manageable, self-contained solutions.

Legal, Commercial, and Operational Challenges in Negotiating Power Purchase Agreements, Firm Supply, and Dual-Use Infrastructure

BTM projects must address a host of legal, commercial, and operational hurdles. From a legal standpoint, regulatory compliance is key. Outside of ERCOT, FERC plays a pivotal role by regulating generator interconnection and surplus power sales. A recent example is the rejection of Talen Energy’s proposal to co-locate an Amazon Web Services data center next to a nuclear plant in Pennsylvania, due to concerns over cost impacts and grid reliability.

ERCOT requires certain nodal protocols be satisfied to authorize connection of the grid with a generation resource that will also serve a co-located BTM load, along with requiring that the energy generated by the project be bought and sold through a retail electric provider and a qualified scheduling entity.

In other states, a generator making BTM retail sales will need to ensure that such retail sales are permissible and that it will not become a utility regulated by the state public service commission. 

A generator making BTM retail sales also needs to ensure that the configuration does not result in it providing any wires services that could potentially subject it to state public service commission regulation.  Also, retail sales by a generator could create a risk to upstream owners because it might subject them to FERC access pursuant to the Public Utility Holding Company Act of 2005. Additionally, BTM projects must meet environmental regulations, particularly emissions rules affecting gas-fired systems. Legal contracts must also address enforceability issues like force majeure, termination rights, and performance guarantees.

These mechanisms help a data center, on the one hand, and the generator, on the other hand, allocate risk for events that may be outside of their control but could drastically affect both projects’ economics.

Commercially, price volatility remains a concern. Fuel-based BTM systems must negotiate fixed or hedged prices in their PPAs to ensure cost stability. Creditworthiness is another obstacle, as energy producers prefer financially strong counterparties—often excluding smaller or emerging data center firms. Firm supply commitments are also critical, given AI workloads’ need for consistent, uninterrupted power when training the models.

On the operational side, dual-use infrastructure—where a generation asset supports both data centers and industrial operations like oilfields—requires careful power allocation. Managing hybrid systems that include gas, renewables, and storage presents interoperability challenges and demands precise energy management.

Scalability is another operational focus, requiring proactive planning for capacity expansions and maintenance. Strategic siting, such as locating data centers near natural gas pipelines and fiber optic networks, as seen in the Texas Critical Data Centers initiative, also enhances long-term viability.

By leveraging natural gas, renewables, and hybrid systems that couple gas and renewable generation sources, developers can convert underused or undervalued natural gas into scalable, cost-effective and reliable power. Long-term offtake and co-location agreements provide the financial and operational certainty needed to support these types of large-scale deployments. While regulatory hurdles remain, particularly around interconnection and the impact on the transmission systems of public utilities, BTM solutions offer a pathway to bypass grid congestion and avoid delays—especially in high-demand regions with good regulatory environments like ERCOT.

As AI continues to drive exponential power consumption, careful negotiation and risk management of power agreements, supply guarantees, and dual-use infrastructure will be essential. Ultimately, BTM projects not only can support sustainable growth but also strengthen grid resilience and unlock new value in energy markets across the nation.

 

About the Author

Bryan Clark

Bryan Clark is partner in the Houston office of Bracewell and works in the midstream oil and gas sector and on energy transition issues. In his work with partner Jared Berg, they have helped electrify the oilfield.

About the Author

Jared Berg

Jared Berg is a partner in the Houston office of Bracewell and works in the midstream oil and gas sector and on energy transition issues. In his work with partner Bryan Clark, they have helped electrify the oilfield.