Part 7 of 13

A Data Center With Its Own Power Plant

By Randy Salars · Source-review edition: September 7, 2026

Before deciding how we feel about the computers, we need to understand the power plant proposed to run them. Site Layer 1 would combine two substantial operations on one site. Producing its own electricity could help the developer secure a dependable supply. It would also place responsibility for continuous fuel delivery, generating equipment, cooling, and emissions close to our communities.

The supplied notice describes 32 Wärtsilä 20V34SG natural-gas engines, each rated at approximately 9.4 megawatts. Multiplying those figures gives 300.8 megawatts of nominal gross generating capacity. The notice describes potential operation around the clock throughout the year. It also identifies emissions controls and ammonia storage. Those details warrant review as a potentially continuous industrial operation, rather than an assumption that the engines would run only during occasional outages. Source: applicant's published notice.

Power is the rate at which electrical energy is produced or used, and a megawatt is a unit of power. A megawatt-hour measures an amount of electricity over time. A plant operating at 100 megawatts for one hour produces 100 megawatt-hours. This distinction matters because an equipment rating does not tell us how much electricity would actually be generated during a year. That depends on operating hours, output levels, maintenance, and outages.

It also matters that 300.8 megawatts is not a confirmed computer load. Some generated electricity would support the plant and facility equipment. The design might also reserve capacity for maintenance or failures. We need a clear account of gross generation, power available after the plant's own needs, total facility demand, and the portion used by the computers. Without that breakdown, comparisons with other data centers can easily become misleading.

There is a practical reason developers consider producing their own electricity. If suitable utility service cannot be obtained on the desired schedule, onsite generation may offer another route. It may also give the operator more control over supply. Whether it actually provides a faster or better solution at this location depends on fuel contracts, equipment deliveries, permitting, construction, and operating costs. The developer should be able to explain those choices without asking residents to take the general business case on faith.

Multiple engines can provide useful flexibility. The manufacturer describes Pearsall Station in Texas as a 24-engine installation using the same 20V34SG engine family. Its account identifies modular operation and closed-circuit radiator cooling. That demonstrates relevant engineering experience with this type of equipment. Pearsall is a different facility serving a different role; its performance and water claims cannot simply be transferred to the proposed New Mexico project. Source: Wärtsilä's Pearsall Station reference.

The manufacturer's own specifications caution that site conditions, fuel, and emissions limits can affect performance and call for project-specific figures. The relevant commitment is therefore guaranteed output for the selected installation, with its actual cooling equipment and local operating conditions. The design should show how many engines normally run and how maintenance occurs while the computers remain online. Source: Wärtsilä 34SG specifications.

Natural gas is the next link in the chain. Proximity to a pipeline does not establish that enough gas can be delivered at the required pressure throughout the year. A meaningful fuel plan should identify the connection, any additional pipeline or compression facilities, service terms, and what happens during interruption. Public officials should know which costs are private obligations and whether any public infrastructure is involved.

Fuel consumption cannot be determined from megawatts alone. The Energy Information Administration explains that the quantity depends on generating efficiency and the fuel's energy content. A “heat rate” expresses the fuel energy needed to produce a unit of electricity; a lower heat rate means less fuel for the same output when measured on the same basis. Expected and maximum hourly and annual gas consumption should follow from the proposed operating schedule. National averages cannot substitute for those calculations. Source: EIA's explanation of fuel used to generate electricity.

Efficiency deserves attention beyond the fuel bill. A generating plant converts part of its fuel energy into electricity; the rest must be accounted for, much of it as heat. EIA distinguishes electricity generated from the amount available after a plant's own power requirements. The community should ask for performance figures that use consistent boundaries, so an attractive gross rating does not obscure the resources required to deliver useful power. Source: EIA's explanation of power-plant efficiency.

The combined facility therefore needs an integrated cooling explanation. The power equipment has heat to manage, and so do the computers using its electricity. A claim about conserving water at the engine plant would not settle water use inside the data center, or vice versa. Engineers should show how the full system performs during the conditions for which it must be designed, including the hottest expected operating periods.

Reliability also reaches beyond the number of engines. If multiple units share a fuel connection, switching equipment, or a control system, the public should ask how failures in those common components are handled. We do not need security-sensitive diagrams to understand whether an interruption would shut down computing, trigger additional equipment, or create demands on public services. The emergency plan should describe responsibilities in terms local responders can use.

Residents may reasonably wonder whether a plant this size could help keep nearby towns powered. That possibility requires actual connections, equipment, operating arrangements, and agreements. It is not an automatic benefit of locating generation in the county. The Department of Energy discusses ways data centers can contribute flexibility to power systems, but realizing those benefits depends on how a facility is connected and operated. No documented commitment reviewed for this proposal guarantees emergency electricity to nearby communities. Source: DOE's discussion of resources and flexibility for data-center electricity demand.

A well-designed private power plant could solve a real business problem. Whether it also serves Grant County depends on its resource demands, effects on neighbors, and financial obligations. I would judge the proposal by the electricity it can reliably deliver under actual conditions and by who carries the consequences when something goes wrong.

What a useful answer would include

The developer and equipment supplier should provide guaranteed local-condition output, fuel consumption, cooling requirements, and a clear account of where the electricity goes. Gas and electric providers should verify connection commitments. The outage plan should assign operational, emergency-response, and payment responsibilities, including failures of equipment shared by multiple engines.

Questions worth asking

• What are the gross generating capacity, dependable net output, total facility demand, and confirmed computer load?
• What operating schedule, maintenance reserve, and hot-weather performance does the design assume?
• What gas supply and delivery agreements are secured, and what new infrastructure and interruption arrangements are required?
• How would the entire power-and-computing facility manage heat, water use, emissions, and shared equipment failures?
• Would the project exchange power with the grid or provide community backup, and what agreements establish the benefits and costs?

AI assisted most of the research and initial drafting under my direction. Sources are linked; corrections are welcome.