Feasibility File · STACK Infrastructure

Stafford Technology Campus

Stafford County · VA

Gated megasite, not a cleared gigawatt campus. Plausible land-use and demand case in the world's deepest data-center market — held below 5 until Dominion service, cost-allocation, water/sewer, and tenant evidence move from announcement to proof.

Verdict

4.8/ 10

Bearing Labs feasibility score

Treat STC as a gated megasite, not as either a dismissed proposal or a cleared gigawatt campus. The project's strengths are real: large controlled acreage, a specific Stafford land-use record, proximity to Northern Virginia demand, Dominion infrastructure context, and a fiscal-benefit narrative with local political appeal. The weaknesses are equally real: the public record does not yet prove firm 1+ GW power delivery, cost-causation protection, a 2027 live date, Stafford-specific capital sufficiency, water/sewer sustainability, or tenant absorption. Score should remain 4.8 until specific named evidence is public: Dominion phase-by-phase service commitments, SCC/PJM cost-allocation treatment for STC-related upgrades, final site/civil and building permits for phase one, a binding water/sewer capacity and CIP-relief package, STC-specific financing allocation, and tenant commitments showing contracted MW by phase.

Where the story diverges

What the operator says

500+ acre campus in the Falmouth area of Stafford County, east of US Route 1 at I-95. Stafford County materials identify a 523.94-acre rezoning case for approximately 5.5 million square feet of data and computer services uses, located around Richmond Highway, State Shop Road, and Eskimo Hill Road in the Falmouth Election District. Planned as four sub-campuses with 19 planned data centers, with zoning capacity reportedly up to 23 buildings / ~5.8M sq ft.

Power architecture: six 300 MW Dominion Energy substations, first-phase service targeted for 2027. Public benefits cited at more than $80 million in annual tax revenue at full buildout, roughly $58 million in county water/sewer capital-improvement relief, hundreds of permanent jobs, and 1,000+ peak construction jobs. STACK's broader financing story includes an approximately $4 billion green-financing package shared across STC and other expansion markets.

What's actually there

STC is a credible hyperscale data-center campus because it combines a very large entitled land position, a Northern Virginia market location, stated Dominion substation architecture, a strong sponsor profile (STACK Infrastructure / IPI Partners), and a plausible public-infrastructure benefit story. Transmission screening around the Falmouth area shows multiple in-service lines within 10 miles, including 500 kV and 230 kV infrastructure owned principally by Virginia Electric & Power Co. The September 2023 rezoning materially improves the baseline.

But the hard feasibility question is not whether Stafford County can host data centers — it is whether a 1+ GW flat load can be energized, paid for, cooled, permitted, occupied, and defended politically on the advertised timetable. The decisive proof package is not public: no executed Dominion service agreement, firm-load letter, PJM/Dominion network-upgrade studies, substation-by-substation in-service dates, customer-specific cost-allocation terms, named anchor tenant, contracted MW, water/sewer service capacity, blowdown handling, or enforceable low-water cooling basis of design. EIA reported PJM expects the Dominion zone to have the largest absolute summer-peak increase from 2026 through 2030, largely from data-center load growth; JLARC concluded growing energy demand is likely to increase costs for other customers; the Virginia SCC's 2025 Dominion order responded by requiring certain large-scale customers to pay at least 85% of contracted distribution and transmission demand and 60% of generation demand. That regulatory response is a sign the risk is real enough to regulate — not a theoretical objection.

Risk 01

Single most decision-relevant exposure

Firm power deliverability plus cost allocation

The project claim is that Dominion Energy can serve 1+ GW through six committed 300 MW substations while Stafford's campus coexists with the broader Northern Virginia load surge — including other Stafford-area hyperscale demand — without destabilizing reliability or shifting material costs to ordinary Dominion customers. That is the claim whose failure collapses the rest of the campus story. A 500-acre rezoned site, a green-financing headline, and strong market demand do not produce a live data-center campus without firm energization. Conversely, a 2027 first-phase live date is not credible unless the public record shows Dominion service documentation, phase-by-phase substation milestones, network-upgrade scope, SCC/PJM treatment of cost allocation, and tenant load ramps tied to the same delivery dates.

Claim Scorecard

Every announced claim, scored.

Each claim assessed independently against supporting and contradicting evidence in the full report.

01
5.5/ 10
Dominion can deliver 1+ GW via six committed 300 MW substations on the project timeline
Meaningful but qualified
Physical-grid proximity and substation narrative support plausibility; firm 2027 deliverability is not public — substation nameplate is not the same as firm service.
5.5/ 10
02
5.0/ 10
Construction starts 2025, first phase live in 2027, and the 19-building buildout progresses
Meaningful but qualified
Zoning foundation is real; downstream permits and utility milestones are missing — the schedule is plausible as a target, not proven as a delivery plan.
5.0/ 10
03
4.5/ 10
STC's share of the ~$4B green financing is sufficient and durable to complete the campus
Preliminary
Sponsor capital access is credible; Stafford-specific allocation and full-campus capital plan are not disclosed — at 25% allocation, ~$1B equals ~$909,000 per MW for a 1.1 GW campus.
4.5/ 10
04
5.0/ 10
Demand is durable for a 1+ GW wholesale campus without a disclosed named anchor tenant
Meaningful but qualified
NoVA demand is real (JLARC reported Dominion forecasting data-center peak demand reaching 9 GW over 10 years); no STC anchor, contracted MW, or take-or-pay demand proof is public.
5.0/ 10
05
4.0/ 10
Cooling water needs can be met without material local impact, and water/sewer CIP relief holds up
Meaningful but qualified
Utility-impact materials exist; peak-day water demand, source, sewer, blowdown, drought-stage rules, and $58M CIP-relief detail are not public — FEMA flood-hazard screening identifies Special Flood Hazard Area polygons near the Falmouth/Eskimo Hill Road area.
4.0/ 10
06
4.0/ 10
STC coexists with broader NoVA load without reliability harm or ratepayer cost shifting
Meaningful but qualified
Controlling claim. Virginia's ratepayer and large-load proceedings show unresolved system risk; no cumulative-load treatment for STC plus other Stafford and NoVA demand is public.
4.0/ 10
07
5.5/ 10
Community and political support remains durable despite Virginia data-center backlash
Meaningful but qualified
Stafford entitlement and fiscal narrative help, but statewide ratepayer, transmission, water, and noise issues are rising — STC has a support base, not yet a durable mitigation package.
5.5/ 10
08
4.5/ 10
More than $80M annual tax revenue and ~$58M CIP relief materialize at full buildout
Preliminary
Gross benefit is plausible at full buildout; no public fiscal model or binding CIP schedule validates it. A one-year delay in the full tax stream reduces the present value of the public benefit by ~$63M at an 8.25% discount rate.
4.5/ 10

Cross-Cutting Risk Factors

The interfaces that decide the project.

01
Power schedule and construction schedule are inseparable
A data center can be graded, shelled, and even partially fit out without the campus being economically live. STC's 2027 first-phase claim depends less on generic construction speed than on Dominion energization. Six substations are an impressive concept, but the schedule turns on engineering, procurement, siting, easements, equipment lead times, network studies, cost allocation, and commissioning.
02
Ratepayer politics can reprice the project even after local approval
The SCC's large-load protections are constructive for public policy, but they also raise the bar for the project. A campus of this size must show who pays for network upgrades, minimum demand commitments, exit fees, collateral, and stranded-asset risk. Without that record, the project remains vulnerable to a statewide 'data centers are raising bills' narrative.
03
Water and sewer claims need the same proof ladder as power claims
The $58M CIP-relief story is a benefit claim, not a capacity proof. The campus needs a public water balance: cooling design, potable versus non-potable share, peak demand, annual consumptive use, sewer acceptance, blowdown treatment, drought operations, and capital responsibility. Without that, water remains a political and environmental risk even if the engineering can ultimately be solved.
04
Merchant demand cuts both ways
A wholesale developer in Northern Virginia benefits from the world's deepest demand pool. The same merchant posture weakens public substantiation when no anchor tenant is disclosed. Under the Feasibility Index methodology, market demand can support plausibility, but it cannot move the claim into the 7–8 band without process-or-proof evidence specific to the project.
05
Fiscal benefits are back-end loaded
The headline tax number is a full-buildout claim. Any delay in energization, absorption, building permits, or utility infrastructure pushes the benefit curve out. The $80M annual benefit at an 8.25% discount rate loses ~$63M in present value from a single-year delay.
06
Reference class is Northern Virginia, Meta Richland Parish, and 'phantom load'
NoVA shows large data-center load in the Dominion zone can be delayed by transmission constraints even when demand is real. Meta Richland Parish shows what a stronger public power case looks like (named customer, filed utility proceedings, customer payments, minimum monthly payments, termination protections). xAI Colossus / Microsoft Quincy show 'phantom load' risk — proposed data-center load can be duplicated across markets, and commercial-readiness standards can materially reduce apparent demand.

Project Profile

Reference data from the public record.

Project: Stafford Technology Campus
Stage tag: Permitting
Sponsor / operator: STACK Infrastructure (IPI Partners-backed)
Location: Falmouth area, Stafford County, VA · Richmond Highway / State Shop Road / Eskimo Hill Road · east of US Route 1 at I-95
Site scale: 523.94-acre rezoning · 19 planned data centers across 4 sub-campuses · up to 23 buildings / ~5.8M sq ft zoning capacity
Zoning: September 2023 rezoning approval · Tax Map Parcels 38-124 and 38-29A · October 2023 data-center zoning framework
Power architecture: Six 300 MW Dominion substations · ~1,100 MW end-state · 1,800 MW gross substation nameplate
Financing context: ~$4B green-financing package shared across STC, Portland, Toronto — STC-specific allocation undisclosed
Public benefits cited: >$80M annual tax revenue at full buildout · ~$58M county water/sewer CIP relief · hundreds of permanent jobs · 1,000+ peak construction jobs
Schedule (Bearing Labs read): First material construction plausible 2026–2027; first live phase more likely 2028+ unless utility delivery and civil approvals are further along than the public record shows
Regulatory context: Virginia SCC 2025 Dominion order requires certain large-scale customers to pay ≥85% of contracted distribution/transmission and 60% of generation demand
PJM context: PJM expects Dominion zone to have the largest absolute summer-peak increase 2026–2030, largely from data-center load growth (EIA)
Scoring arithmetic: Claim mean 4.8125 → 4.8 · controlling claim 4.0 · cap 5.5 (does not bind)
Evidence as of: May 29, 2026

Bearing Labs · Feasibility Brief