AI QUANTUM PARK

FINAL CAMPUS LAND ALLOCATION & CLASS A COST ESTIMATE

​

Virginia, USA | 22,000-acre holding | 2,000-acre Phase I campus footprint | 8 districts

 

 

​

 

 

​​

 

Executive conclusion. The original equal 166.67-acre split was mathematically clean but not operationally optimal. This revision reallocates land according to real estate intensity, safety/setback needs, logistics, expansion requirements, and the relationship between two 900 MW generation campus and a large AI/HPC data center. The resulting recommended Phase I planning budget is approximately $18.77 billion, excluding land acquisition, financing, FF&E;, active AI/GPU hardware, and major off-site utility/network extensions.

 

1. Recommended 2,000-Acre Land Allocation

District Acres % of Campus Planning Logic

Aerospace & Drone 450 22.5% Large hangars, test areas, secure buffers and future expansion

AI Data Center 300 15.0% Supports ~500 MW planning IT load with expansion headroom

Advanced Manufacturing & Robotics    400    20.0%                   Large production buildings, robotics lines, staging and suppliers

 

Power Generation 250 12.5% 819 MW plant, gas/electrical infrastructure, safety and acoustic buffers

Commercial Innovation 120 6.0% Dense offices, R&D, incubators, conference and amenity uses

Industrial & Logistics 250 12.5% Warehouses, truck courts, distribution and service infrastructure

Residential 140 7.0% Private apartments/residences, landscape buffers, pond and amenities

Hospitality & Entertainment 90 4.5% Hotels, food, entertainment, event and visitor uses

TOTAL 2,000 100.0% Phase I campus footprint

 

Why this is stronger: the two most land-intensive employment uses remain dominant, but data centers, power, and logistics receive substantially more room than under

the equal-split model. Residential and hospitality remain meaningful but compact, preserving privacy and avoiding unnecessary consumption of the industrial/technology

core.

Important: 300 acres does not mean 500 MW is guaranteed. It is a planning basis. Power density, grading, water/cooling strategy, substations, setbacks, environmental constraints, fiber routing and utility interconnection will determine final usable capacity.

 

2. Revised Class A Capital Cost Model

# Department Acres GFA / Capacity Benchmark Design A - Design B - Construction C - Total

1 Aerospace & Drone Campus 450 2,352,240 $550/sf 10% $129.4M $1,293.7M $1,423.1M

2 AI Data Center Campus 300 500 MW planning IT $12.43M/MW* 9% $559.4M $6,215.0M $6,774.4M

3 Advanced Manufacturing & Robotics 400 5,227,200 $425/sf 8% $177.7M $2,221.6M $2,399.3M

4 Power Generation Campus 250 819 MW generation $2.20M/MW 7% $126.1M $1,801.8M $1,927.9M

5 Commercial Innovation District 120 1,829,520 $550/sf 10% $100.6M $1,006.2M $1,106.9M

6 Industrial & Logistics District 250 4,900,500 $150/sf 6% $44.1M $735.1M $779.2M

7 Residential District 140 1,829,520 $300/sf 8% $43.9M $548.9M $592.8M

8 Hospitality & Entertainment District 90 980,100 $625/sf 10% $61.3M $612.6M $673.8M

Base design (A) $1.242B

Base construction / EPC-GC (B) $14.435B

Vertical + core infrastructure subtotal $15.677B

Internal campus backbone allowance (7.5% of B) $1.083B

Subtotal before contingency $16.760B

Program contingency (12%) $2.011B

RECOMMENDED PHASE I PLANNING BUDGET $18.771B

* Data center benchmark uses JLL's 2026 average shell/core construction cost of $11.3M/MW plus a 10% liquid-cooling premium, producing $12.43M/MW. Active IT/GPU hardware is excluded.

 

3. Key Corrections & Technical Basis

A. Power generation corrected to 819 MW

The prior 140 MW assumption was inconsistent with the project program. This model uses 819 MW. For planning, it applies $2.20M/MW ($2,200/kW), which is consistent

with current PJM-region forward combined-cycle project indications around $2,100-$2,300/kW for late-2020s/early-2030s projects. This yields $1.802B construction/EPC

plus a 7% design allowance.

B. AI data center changed from square-foot pricing to MW pricing

The model assumes 500 MW of critical IT capacity as a prudent Phase I planning target within the 300-acre data-center district. At JLL's 2026 $11.3M/MW average

shell/core benchmark, with a 10% liquid-cooling premium, construction is approximately $6.215B. A 9% design allowance produces $6.774B before campus-wide

infrastructure and contingency.

C. Active AI compute is deliberately separated

JLL notes that tenant technology fit-out for AI infrastructure can reach as much as $25M/MW. At 500 MW, that represents up to $12.5B of additional tenant/owner

technology investment. It is not included in the $18.77B campus development budget because GPU/IT equipment ownership and refresh cycles are fundamentally different

from real-estate construction.

D. Internal campus backbone added

A 2,000-acre multi-district campus requires major internal roads, grading, drainage, security, fiber, shared substations/distribution, landscape systems and utility corridors. A

conceptual 7.5% allowance on base construction has therefore been added. This is a planning allowance, not a substitute for civil, utility and geotechnical design.

E. Virginia location acknowledged

The project is in Virginia. The vertical building benchmarks remain conceptual national/Class A assumptions because the exact county, labor market, topography,

geotechnical conditions, gas interconnection point and utility scope have not been provided. Those items can move actual bids materially.

 

4. Final Planning Position

Recommended figure for presentations: Approximately $18.8 billion Phase I campus development budget for the 2,000-acre program, based on the corrected 819

MW power plant and a 500 MW AI data-center planning load.

 

Do not describe $18.8B as a guaranteed bid price or appraised value. It is a concept-level capital planning estimate. A defensible investment-grade budget requires site

survey, topography, geotechnical report, gas pipeline pressure/capacity, electrical one-line and interconnection study, water/cooling strategy, environmental constraints,

building program, phasing and contractor pricing.

 

Excluded from $18.8B

Land acquisition/value; financing and interest carry; active GPUs/servers/network electronics; tenant-specific FF&E; major off-site gas, transmission or water extensions

beyond the campus backbone; extraordinary environmental remediation; taxes/incentives; escalation beyond the benchmark basis; and operating reserves.

 

Potential fully equipped AI exposure

If the owner also funds a maximum-order AI technology fit-out at up to $25M/MW for 500 MW, the additional technology investment could be as high as approximately

$12.5B, implying a combined real-estate/infrastructure + technology exposure of roughly $31.3B. This is an upper planning scenario, not a recommendation to budget the

full amount without a defined compute architecture.

 

Selected 2026 benchmark sources

• JLL, 2026 Global Data Center Outlook: 2026 average shell/core cost $11.3M/MW; liquid-cooled facilities +10%; AI tech fit-out can reach $25M/MW.

• PJM, Powering Reliability Through Market Design (May 2026): forward combined-cycle projects commonly $2,100-$2,300/kW in PJM for later CODs.

• U.S. EIA, Electric Generator Construction Costs (2026 release): historical installed natural-gas generator cost data used as a cross-check.

• Enverus Intelligence Research (Apr. 2026): 2.0 MW per buildable acre assumption used in its national data-center parcel analysis as a land-density cross-check.

Prepared as a conceptual R&D; / master-planning estimate - August 2026