Geothermal Powered Data Centers: BTM Opportunities for Developers
- Mehrbano Asim
- 3 hours ago
- 5 min read
Data center power demand isn't slowing down, and the grid isn't speeding up. For developers chasing gigawatt-scale AI and hyperscale campuses, the core bottleneck is getting power. Most developers still want grid power first because it's familiar, it's regulated, and it's (usually) cheaper on paper. But interconnection queues in many regions now stretch past 10 years, and no amount of capital can buy your way to the front of that line. That gap is why behind-the-meter (BTM) power has gone from a hedge to a core site selection strategy, and why geothermal is emerging as one of the most interesting BTM options on the table.
The State of the Interconnection Queue Problem
Every regional transmission organization is dealing with the same story: a backlog of generation and large-load interconnection requests that vastly outpaces the grid's ability to process, study, and build for them. PJM alone has reported over 200GW of new grid connection requests moving through its reformed interconnection process - a number that illustrates just how much load is competing for the same finite grid capacity. For a data center developer, a 7-10 year wait isn't a delay; it's a dealbreaker. Capital markets, hyperscale tenants, and AI training timelines don't operate on decade-long horizons.
That's the core tension driving the current wave of site selection strategy: developers need power now, but the grid is built for a slower world. By deploying generation assets directly on-site or adjacent to facility boundaries, developers bypass grid queue friction, eliminate transmission and distribution (T&D) losses, and secure firm power on compressed timelines.
BTM Power Options Data Center Developers Are Already Using
BTM isn't a single technology, and the developers moving fastest are the ones building a power stack rather than betting on one source. Natural gas, solar, battery storage systems (BESS), are among the most common BTM solutions today, but geothermal is emerging as an alternative.
Natural Gas: Behind-the-meter gas generation remains the fastest path to megawatts for many AI-scale campuses, particularly where a site sits near existing pipeline infrastructure with available offtake capacity. Natural gas is dispatchable, well understood by financiers, and can be deployed on a faster timeline measured provided the site has the pipeline proximity and delivery point access to support it. However, developers face mounting regulatory barriers, air quality permitting delays, fuel pipeline capacity bottlenecks, and corporate Scope 1 decarbonization mandates.
Solar: On-site or co-located solar has become a standard component of BTM power stacks, especially when paired with storage. It's cheap to deploy at scale and increasingly bankable, but daylight-only generation means it rarely stands alone for a facility that needs to run 24/7.
Battery Energy Storage Systems (BESS): BESS is the piece that makes intermittent BTM sources (solar in particularly viable for continuous data center loads. Storage extends the value of solar generation beyond daylight hours and provides the buffer that makes a hybrid BTM stack reliable enough for mission-critical operations.
Geothermal: This is the newest entrant to the BTM conversation, and it's moving from experimental to mainstream faster than most developers expected. Unlike solar, geothermal is dispatchable and runs around the clock regardless of weather. Unlike gas, it doesn't carry fuel price volatility or combustion emissions. For developers evaluating long-term power contracts in geothermal-favorable regions, it's starting to look less like a novelty and more like a genuine baseload option.
This shift is exactly what LandGate's Jacob Reed and special guest Gordon Dolven, Director of Americas Data Center Research at CBRE, unpacked in LandGate’s recent webinar, Greenfield Data Center Development: Where the Map Is Moving and Why. The conversation traces how the top data center markets have changed since 2020 and why power availability is now the primary force reshaping where developers choose to build.
As grid capacity gets more constrained, Dolven points to Nevada as a leading example of developers getting creative with behind-the-meter solutions, including nuclear, natural gas, and geothermal, to keep projects moving on timeline.
Case Study: Amazon’s 700 MW Nevada Deal Signposts Geothermal's Rise
The shift toward geothermal-backed hyperscale infrastructure is moving into real-world commercial deployment. In a landmark deal in Nevada, Amazon Web Services (AWS) partnered with utility NV Energy to secure 700 MW of new carbon-free energy tailored to support future hyperscale data center operations in the Reno market.
The transaction highlights how hyperscalers are blending next-gen baseload geothermal with solar and storage to achieve firm, 24/7 clean power:
100 MW Firm Geothermal: Contracted under a 20-year Power Purchase Agreement (PPA) with geothermal innovator Zanskar, leveraging predictive AI modeling and advanced subsurface exploration to unlock greenfield geothermal capacity by 2030.
600 MW Solar + 600 MW BESS: Developed by Primergy to capture peak daytime solar generation, using long-duration BESS to balance intra-day ramps and deliver grid capacity support.
This structure marks Amazon’s first data center deployment leveraging geothermal energy. It provides a clear industry template: combining continuous geothermal energy with solar-plus-storage allows developers to bypass grid constraints while hitting aggressive 24/7 Carbon-Free Energy (CFE) operational metrics.
Amazon isn't alone. Google has also signed a geothermal agreement with NV Energy, and Meta has signed two geothermal deals, with XGS and Sage Geosystems respectively. When four of the largest data center developers in the world are independently underwriting geothermal PPAs in the same 18-month window, that's not an experiment - it's a category forming.
What Geothermal Data Centers Mean for Site Selection
The Amazon deal is instructive beyond the headline megawatt figures. It shows that geothermal viability is something that developers should be considering. Locating the optimal site for a geothermal or hybrid BTM data center requires multi-layered geospatial, geological, and infrastructure analysis. Overlaying these variables manually across disparate datasets creates friction and delays site control.
LandGate’s Data Center Siting Tools centralize these critical spatial layers into a single geospatial analytical interface:
Geothermal Power Plant & Subsurface Data Layers: Access mapped data on existing operational geothermal power plants alongside critical geological indicators - including mapped thermal springs, active and historical thermal wells, and granular temperature-at-depth modeling across key subsurface strata.
Electrical Infrastructure & Offtake Capacity: Analyze high-voltage transmission lines, substation capacity, PNode pricing points, and regional ISO/RTO interconnect bottlenecks to target ideal off-grid or hybrid interconnection zones.
Multi-Vector Energy Layers: Evaluate co-location suitability by overlapping natural gas pipelines, solar farms, existing BESS deployments, and fiber optic routes (long-haul, regional, dark fiber lines).
Parcel-Level Buildability & Exclusions: Run automated buildable area analyses that calculate parcel acreage, slope, wetland/floodplain constraints, environmental easements, local zoning rules, and direct landowner contact data for fast option agreements.
Layering these data sets against parcel-level site data turns power sourcing from a downstream surprise into an upfront filter, eliminating sites that can't support the power strategy before a developer spends months on entitlement and negotiations.
Don't let 10-year grid queue delays halt your development timeline. By leveraging Behind-the-Meter hybrid power strategies anchored by next-generation firm geothermal energy, data center developers can secure immediate, scalable power solutions. LandGate provides the geospatial intelligence, subsurface thermal data, and site selection automation necessary to identify, evaluate, and acquire optimal BTM data center sites.
