17-page report
Feasibility Prospectus
Engineering, cooling and water, onsite power options, community design, zoning and permitting, risks, and decision gates.
Feasibility • Community value • Project materials
A single access point for the feasibility study, cost scenarios, Virginia–Texas investor research, global underground-contractor screening, presentation, and four illustrative park concepts.
Project owner: Masters Energy Inc. USA, a Virginia stock corporation.
Concept-level U.S. study • No site or jurisdiction selected • Renderings are illustrative
Project brief
The study supports early screening and discussion. It is not a site-specific design, permit determination, cost estimate, or investment recommendation.
17-page report
Engineering, cooling and water, onsite power options, community design, zoning and permitting, risks, and decision gates.
Stakeholder presentation
A concise presentation of the feasibility verdict, capital and energy economics, regulatory path, and park concepts.
Regional investor update
An 8-page, 47-source diligence addendum for a hypothetical 20 MW critical IT load. Market metrics use different scopes and dates; this is not a parcel-specific approval, power commitment, or investment recommendation.
Market signals, county and city land-use pathways, tax incentives, power delivery, underground site risks, and investment gates. Updated November 8, 2026.
Fundraising strategy
For this early, site-unspecified 20 MW concept, separate technology-company capital from the much larger capital required to build a facility. Treat named funds below as thesis-fit examples, not as confirmed interest or available capital.
Venture capital
Venture capital fits only if the team owns a scalable product such as a proprietary underground conversion/design system, validated life-safety and cooling package, site-screening software, or deployable power-and-controls technology. The pitch must show a paying customer, defensibility, repeatable revenue, and a pilot with measurable cost, schedule, or energy results.
In Virginia, Virginia Venture Partners publishes pre-seed investments up to $250,000 and seed investments of $500,000–$1 million, subject to company stage, traction and round requirements. It may fit an eligible Virginia technology startup; it is not construction finance for a site SPV.
For a true energy-technology startup in Texas, Energy Transition Ventures publicly lists distributed generation, smart grid, storage, demand flexibility and enabling technology as focus areas. That makes it a possible thesis-fit conversation for a defined product, not proof of interest or current capital availability.
Energy infrastructure capital
Asset investors underwrite site rights, permitted use, utility power by phase and date, tenant credit/offtake, independent cost and schedule, and downside protection—not a rendering or a statewide pipeline number. Consider a developer, energy partner, or predevelopment sponsor first. Seek larger asset-level equity or project debt only when the evidence is bankable for that instrument.
Stonepeak and DigitalBridge publish data-center/digital-infrastructure mandates. Treat them as examples of later-stage mandate fit, not assumed local investors for this unselected project. Verify current fund, geography, scale, development-stage appetite and partner requirements before approaching.
Keep an onsite-plant or microgrid case separate: show fuel, interconnection, emissions permits, dispatch limits, redundancy, capital cost, and who bears power-price and curtailment risk. Emergency backup equipment is not unrestricted prime power.
Regional proof points
Use market evidence to establish why a location deserves diligence. Do not turn market reports into claims about a specific site's entitlement, energy capacity, lease, or return.
Virginia
CBRE reported 0.5% Northern Virginia colocation vacancy in H2 2025 and difficulty securing contiguous 15–20 MW blocks. Date and label those as market signals, not a tenant or power commitment.
Name the county and zoning district early: Loudoun, Fairfax and Prince William have distinct pathways. The illustrative underground PUE of 1.30–1.45 turns 20 MW critical IT into roughly 26–29 MW of facility input. Confirm how the utility's 25 MW GS-5 tariff threshold effective January 1, 2027 applies to contract demand, auxiliaries and phasing; model the current electricity-consumption tax and verify exemption eligibility rather than presenting tax benefits as base-case proceeds. [Dominion] [SCC tariff] [Virginia Tax]
Texas
JLL reported 6.5 GW of Texas data-center capacity under construction at year-end 2025. Describe this as pipeline, not operating supply. Map the candidate site to the city, ETJ, county and plat authority, then identify its actual utility/TSP and study path.
The generally cited ERCOT/PUCT large-load framework uses a 75 MW single-site threshold. A 20 MW IT project can still face utility and system studies below that threshold; forecast campus growth and aggregation. Underwrite Texas tax incentives only after eligibility and procurement sequence are confirmed. For either state, diligence water, flood, cooling, air permits for generation, fire/life safety, and the community park's security and maintenance plan.
Suggested sequence
Decide whether capital is for PlatformCo product validation, ProjectCo predevelopment, or both in separate vehicles. Do not ask a VC to fund excavation.
Secure an option, screen geology/water/flood, meet the building official, and obtain written utility feedback with facility load and dates.
VC: one-page product teaser and pilot milestones. Infrastructure: partner referral and a candid asset teaser with unresolved gates disclosed.
Use early sponsor or non-dilutive capital for diligence; seek larger asset capital after site, power, tenant, cost and entitlement evidence harden.
First VC conversation
“We are validating a repeatable [product or system] for data-center developers that is designed to improve [measurable bottleneck]. This round funds [pilot and proof milestones]; it is not a request for venture capital to bankroll one building.”
Fill brackets only with validated claims. Include product, customer, IP, per-deployment economics, repeatable market and use of funds.
First infrastructure conversation
“We are screening a phased 20 MW critical-IT project in [named jurisdiction]. Here is the documented status of site control, land use, utility service, tenant, engineering and cost. We are seeking [development partner or capital] for [specific next gate].”
Do not imply a site, tenant, power contract, permit, budget or return exists unless it is documented.
The hub's current CapEx ranges are screening estimates, not bids; underground premiums are analyst assumptions, and equipment exclusions remain. Reconcile an independent surface comparator, per-MW sources and uses, PUE, power price, taxes, water, insurance, contingency, and lease evidence before presenting a return case. Download the full source-linked strategy with the evidence notes.
Additional evidence: JLARC, Virginia data-center inventory and power estimates; VEDP, Virginia data-center equipment exemption; Texas Comptroller, data-center sales-tax program. Fund examples do not imply current capital availability or investment interest.
Geography-open contractor screen
A sourced screen of U.S. and international subterranean contractors. Profiles distinguish mined caverns and shafts from TBM tunneling, geotechnical enabling works, and standard data-center fit-out; no firm is presented as selected or prequalified.
Public project evidence, likely package roles, capability gaps, official contact paths, and RFI questions. Companies are not excluded based on headquarters or country.
Illustrative economics
Change the IT load and power price to recalculate the screening scenario. Values scale linearly for illustration; real project estimates require site and design inputs.
Modelled as a continuous load for 8,760 hours/year.
Default: EIA 2026 year-to-date industrial average; your actual tariff may differ.
Power usage effectiveness (PUE) is facility energy divided by IT energy. The modeled ranges are assumptions—not measured operating data for underground sites.
Ground-coupled cooling screen
Stable shallow-ground temperatures can support heat exchange or cold thermal storage, but a buried facility does not automatically save cooling energy. This screen keeps measured cooling electricity and assumed savings separate from the main PUE estimates.
Typical relatively stable ground temperature at about 30 ft (9 m); local site conditions vary.
Recommended air-inlet range for data-processing equipment classes A1–A4; ground and server-air readings are different measurement points.
USGS reports average geothermal gradients around 25°C/km in the eastern U.S. and 34°C/km in the West; actual geology and depth matter.
Enter metered or engineered cooling-system electricity only—not total facility energy. Leave 0 until measured or estimated by the design team.
User scenario only. Default is 0%; change only to test an explicit engineering assumption.
Research signal, not a 20 MW forecast: NLR's 2026 modeled Virginia 1-GW Cold UTES case projected about 70% lower annual electricity cost for cooling and a 200-MW lower peak cooling-electric load. Cold UTES stores cold underground during off-peak periods and dispatches it during peak periods; the modeled cost reduction is not the same as a 70% reduction in annual kWh, and it is not a measured result for this project.
For a continuously operating 20 MW IT load, the IT equipment alone uses about 175.2 GWh/year, with nearly all of that electrical input ultimately becoming heat that must be rejected. Ground-loop sizing, heat recharge/rejection, pumps, backup cooling, and hourly tariffs require site-specific engineering. The sensitivity above excludes demand charges, taxes, water, and capital costs.
Sources: DOE, Geothermal Heat Pumps; ASHRAE, 2021 Equipment Thermal Guidelines reference card; USGS, Geothermal Gradients in the Conterminous United States; NLR, Cold UTES data-center model results; NLR Phase 1 Cold UTES report.
Optional onsite-power sensitivity
BWX Technologies (BWXT), headquartered in Lynchburg, Virginia, is the prime builder and integration lead for the Project Pele mobile microreactor prototype. BWXT's selected demonstration unit is designed for 1.5 MWe and uses HALEU/TRISO fuel; a broader federal prototype design envelope discussed in the environmental review is 1–5 MWe. BWXT's current schedule describes the four-container prototype being tested at Idaho National Laboratory, with power operation targeted for 2028. That is a government demonstration, not evidence of a commercially available reactor approved for this underground data-center site.
Optional independent explainer: Apex Builds video on Project Pele, portable reactors, and HALEU/TRISO. This is third-party material, not an official BWXT, DOE, DoD, or NRC source. Its broader civilian-use discussion is forward-looking; claims about commercial availability, performance, price, licensing, and safety are not used as verified project facts or calculator inputs.
Nameplate electric output stated by BWXT for the current Project Pele design.
BWXT identifies Lynchburg manufacturing work; the government prototype is slated for demonstration at Idaho National Laboratory.
Transportable system described in four 20-foot containers; TRISO fuel uses multiple protective coatings but is not risk-free.
Scenario control only. A value of 1 represents one 1.5 MWe prototype-sized unit; it does not imply a commercial unit is available.
Default is 0%, so the hub credits no nuclear energy or savings. Change this only to explore an explicitly hypothetical operating case; BWXT has not published a commercial fleet capacity factor.
Leave blank while unknown. Enter a sourced or clearly labeled user assumption that annualizes CapEx and includes fuel, staffing, maintenance, security, insurance, licensing, waste/fuel disposition, and decommissioning. Blank is not zero cost.
What pricing is public? BWXT's 2022 issuer release described an approximately $300 million cost-type contract, depending on options, to design, manufacture, and deliver a first-of-a-kind government prototype. That amount is not a commercial unit price and must not be divided by output or treated as repeat-unit CapEx. I did not verify a public commercial purchase price, site-installed cost, vendor O&M rate, or fuel/security cost. The calculator therefore leaves annual cost blank until you enter an explicit assumption. Fuel availability is also a diligence risk: GAO reported in 2026 that near-term HALEU supply may not meet demand.
Potential value—not a turnkey solution. If a future product is commercially available, licensed, and economic, firm on-site generation could diversify the supply mix and support resilience; Project Pele's transportable format is a logistics feature. These are conditional possibilities, not established benefits for this site. One 1.5 MWe module cannot make a 20 MW IT campus self-contained, and the calculator assumes no grid-displacement, reliability, emissions, or lifecycle-cost benefit.
What it means for the 20 MW concept. At the current calculator's 20 MW critical-IT default and PUE range of 1.30–1.45, the modeled facility load is 26–29 MW. One 1.5 MWe module equals about 5.2–5.8% of that nameplate load; 18–20 modules would match it on arithmetic alone, before reserves, auxiliary loads, derating, maintenance, outages, UPS, or backup. Nuclear generation does not eliminate data-center cooling: reactor and server waste heat still need engineered rejection. This screen does not assume a reactor can be placed inside the cavern; a separate, accessible, site-engineered power facility is only a concept for evaluation.
Safety and approvals. The Project Pele test is non-commercial and proceeds under DOE authorization at INL; that is not a commercial site license. The NRC states that commercial reactor design, siting, construction, and operation involve licensing and regulatory oversight, including public involvement. TRISO's coated-particle design is a safety feature, not a promise of zero risk. A real installation would require its own regulator/jurisdiction determination, environmental and emergency planning, security, fuel transport and disposition, insurance, staffing, and decommissioning analysis. The two outputs marked with an asterisk are simple energy-value sensitivities before those costs and are not savings, LCOE, project returns, or an investment forecast.
Sources: BWXT, Project Pele status, output, format and schedule; BWXT, 2025 fuel-manufacturing update; Idaho National Laboratory, first TRISO fuel delivery; DoD environmental review, prototype envelope and demonstration pathway; BWXT issuer release, 2022 prototype contract; GAO, 2026 HALEU supply review; BWXT, TRISO fuel design; NRC, commercial new-reactor regulation; EIA, electricity-price reference used in the calculator.
CapEx starts from Cushman & Wakefield’s 2026 North American greenfield average of $17.6M/MW, excluding chips and GPUs; confirm its MW denominator matches critical IT load. Underground uplifts (15–50% for mine/cavern reuse; 40–100% for new excavation) are analyst screening assumptions, not published underground benchmarks or bids. Base CapEx and energy-cost cases exclude onsite generation; the separate Project Pele sensitivity above is not a commercial price quote. Energy expense is power only, before demand charges, taxes, staffing, maintenance, insurance, water, and financing. Underground systems may perform better or worse depending on geology and design.
Sources: Cushman & Wakefield, 2026 Data Center Development Cost Guide; U.S. EIA, Electric Power Monthly (2026 YTD industrial average: 9.09¢/kWh, preliminary).
Recommended posture
Screen geology, groundwater, flood exposure, access, utilities, fiber and permitting before concept design.
Study an islandable microgrid as an option—not a default. Plant permits, firm fuel and dispatch limits matter.
Design park access, truck separation, noise controls, load limits and maintenance obligations together.
Illustrative concepts
Select a concept image to enlarge it. Each rendering depicts a hypothetical site, not an approved plan.
Flexible gathering space with raised planters and stewardship.
Download renderingOne file for everything
The refreshed ZIP includes the original feasibility report and park renderings, Virginia–Texas investor materials, the fundraising strategy, the geography-open subterranean contractor report, and the source-cited BWXT Project Pele microreactor screening addendum.
Keep a local copy for investor review, offline access, or project archive.
Download the source-cited BWXT Project Pele addendum (Markdown)