1 GW Data Center Campus
Capital Investment Analysis
This analysis models a 1,000,000,000-watt (1 GW) data center energy campus with a total capital budget of $1,200,000,000 across three rack density configurations. Each scenario examines the trade-offs between power density, physical footprint, and all-in project cost — providing infrastructure planners and financial decision-makers with a clear framework for evaluating build strategy at gigawatt scale.
The three scenarios span 200,000 W rack power density at 10,000, 15,000, and 20,000 racks respectively. Each configuration produces a distinct campus footprint, shell cost, and total project cost. The analysis isolates the vanilla powered shell cost (excluding Energy and IT systems) to surface the pure real estate and structural cost differential between configurations.
Scenario Overview: Three Density Configurations
At a fixed 200,000 W per rack across all three scenarios, the total rack count is the primary variable driving campus scale. Higher rack counts reduce power density per square foot, expanding the physical footprint and cascading through every cost line. The table below captures the core parameters of each configuration side by side.
The range in campus footprint is striking: Scenario A demands 680,272 SF — more than four times the footprint of Scenario B at 158,730 SF. Scenario C lands in between at 238,095 SF. This footprint spread has direct implications for land acquisition, permitting timelines, utility interconnect complexity, and civil infrastructure cost.
Powered Shell Cost: Vanilla Structure Only
The vanilla powered shell cost — explicitly excluding Energy systems and IT equipment — isolates the structural, civil, and building envelope investment required to house the compute at each density tier. This metric is critical for understanding how architectural density decisions translate into base construction cost before any mechanical, electrical, or IT overlay is applied.
Scenario A — 10,000 Racks
$238,095,238
680,272 SF campus shell at the largest footprint. Lowest cost per rack for shell construction but highest absolute land and civil exposure.
Scenario B — 15,000 Racks
$79,365,079
158,730 SF campus — the most compact configuration. Lowest absolute shell cost driven by the smallest physical footprint of the three scenarios.
Scenario C — 20,000 Racks
$166,666,667
238,095 SF campus. Moderate shell cost reflecting a mid-range footprint, with the highest rack count yielding the greatest compute density per dollar of shell investment.
Scenario B delivers the lowest vanilla shell cost by a significant margin — approximately 67% less than Scenario A and 52% less than Scenario C. However, shell cost is only one component of total project economics. The decision cannot be made on structural cost alone.
Total Project Cost: All-In Per Configuration
Adding the $1,200,000,000 energy and IT system budget to the vanilla shell cost produces the all-in total project cost for each scenario. Despite the wide variation in shell costs, the total project costs converge — demonstrating that at gigawatt scale, the energy and IT systems dominate the capital stack, and shell cost differentials represent a smaller fraction of total investment than they might appear in isolation.
Scenario B produces the lowest total project cost at $1,429,365,079, driven by its dramatically smaller physical footprint and correspondingly low shell cost. Scenario C carries the highest total cost at $1,566,666,667, despite having the highest rack count — reflecting elevated shell cost per SF at that density tier. Scenario A lands at $1,538,095,238, with total cost elevated by its large campus footprint.
Cost Decomposition: Shell vs. Energy & IT Systems
Understanding the proportion of total project cost attributable to the powered shell versus the energy and IT systems is essential for identifying where cost optimization levers exist. Across all three scenarios, the $1.2B energy and IT budget represents the dominant cost driver — yet the shell cost differential of ~$159M between the cheapest and most expensive configuration is material at this project scale and should not be dismissed.
The stacked view confirms that energy and IT systems consume between 84% and 95% of total project budget across all scenarios. Shell cost as a share of total investment ranges from approximately 5% (Scenario B) to 15% (Scenario A). This means decisions about rack density and campus layout — while consequential — are secondary in financial magnitude to equipment and energy procurement strategy.
Footprint & Density: The Physical Trade-off
Campus square footage is one of the most consequential planning variables at gigawatt scale. A larger footprint demands more land, longer utility runs, greater civil and grading cost, more complex permitting, and extended construction timelines. The relationship between rack density and physical footprint is nonlinear and highly sensitive to the racks-per-1,000-SF assumption embedded in each scenario.
The 4.3x difference in campus footprint between Scenario B (158,730 SF) and Scenario A (680,272 SF) is the single most dramatic variance in the dataset. For sites in land-constrained markets or regions with complex environmental review processes, Scenario B's compact form factor may confer substantial schedule and permitting advantages that do not appear in the raw capital cost figures. Conversely, markets with abundant land and low permitting friction may favor Scenario A's more distributed layout if operational flexibility is valued.
Land-Constrained Markets
Scenario B (158,730 SF) is strongly preferred. Minimal footprint reduces land cost, permitting risk, and utility infrastructure requirements.
Land-Abundant Markets
Scenario A (680,272 SF) becomes viable when land is cheap and permitting is streamlined. Larger floor plates can offer operational flexibility and future expansion optionality.
Decision Framework: Selecting the Optimal Configuration
No single configuration dominates across all decision criteria. The optimal scenario depends on the interaction of site constraints, operational objectives, market conditions, and risk tolerance. The framework below maps each scenario to its primary use case and highlights the conditions under which each configuration creates the most value.
Scenario B — 15,000 Racks: Best for Capital Efficiency
Lowest total project cost at $1,429,365,079 and smallest campus footprint at 158,730 SF. Optimal for land-constrained sites, markets with high real estate costs, or programs prioritizing fastest time-to-power. The shell cost savings of $159M versus Scenario A are meaningful and can be redeployed into energy or IT systems. However, the highest racks-per-SF assumption (6,667) requires validation against specific cooling and power distribution architecture.
Scenario C — 20,000 Racks: Best for Compute Density
Highest rack count at 20,000 racks with a moderate footprint of 238,095 SF. Total project cost of $1,566,666,667 is the highest in the set, but also delivers the greatest raw compute capacity. Best suited for workloads requiring maximum parallelism — large-scale AI training, HPC, or multi-tenant hyperscale deployments where per-rack revenue justifies the premium total investment.
Scenario A — 10,000 Racks: Best for Operational Flexibility
Largest campus at 680,272 SF with a total cost of $1,538,095,238. The expanded footprint provides the most room for mechanical infrastructure, airside economization, redundant power paths, and future build-out phases. Best suited for operators who expect demand growth, require extensive white space flexibility, or are building a multi-phase campus where future expansion is already planned into the site master plan.
In all cases, the $1.2B energy and IT systems allocation should be treated as the primary optimization target. Shell cost differentials, while significant in absolute terms, are a secondary lever. The configuration decision should be made jointly by infrastructure planners (physical/operational fit) and financial decision-makers (capital deployment efficiency and IRR modeling), with the full project cost stack — including energy procurement, interconnect costs, and construction contingency — modeled for each scenario before commitment.
Key Metrics Summary
1GW
Total Campus Power
1,000,000,000 watts across all three scenarios
$1.2B
Energy & IT Budget
Fixed capital allocation for energy systems and IT infrastructure
200KW
Per-Rack Power
200,000 watts per rack — consistent across all configurations
4.3x
Footprint Spread
Range between smallest (158K SF) and largest (680K SF) campus footprint
$137M
Total Cost Range
Difference between lowest ($1.429B) and highest ($1.567B) all-in project cost
$79M
Lowest Shell Cost
Scenario B vanilla powered shell — best capital efficiency for structure

All costs represent vanilla powered shell only (excluding Energy and IT systems) unless stated as total project cost. Energy and IT systems are modeled as a fixed $1,200,000,000 allocation applied uniformly across all three scenarios. Actual costs will vary based on site conditions, cooling architecture, utility interconnect requirements, and construction market conditions.