OperationalNo goprofile from priorscurated

Loudoun County cluster

Multiple operators · Ashburn, Virginia, United States · 5,000 MW

Densest data center market in the world; interconnection queue and rate pressure are active policy issues.

Supply as reported: Grid, PJM

loudoun.gov

Open in the simulation

Build cost

$172.1B

Months to live

149

gated by 32 construction phases

EBITDA / yr

$49.6B

Payback

3.5 yrs

IRR

12.4%

CO₂ / yr

11,188 kt

320 g/kWh

Jobs

2,000 / 3,635

build / permanent

Composite

E 68 · Env 30 · C 70

No goon “Baseline as reported” (100% grid) · Oct 30, 2026 · meta/muse-wonder-1.0

The verdict is NO_GO for the 4,000 MW, all-grid Loudoun County proposal, and that verdict stands. Economics are not the blocker: a 12% IRR clears the 8% floor, payback is under four years against a seven-year ceiling, and net present value is about $18B on roughly $172B of capital spending. The blockers are an environment score of 30 against a 45 floor, emissions of about 11 Mt a year, a grid draw of 1,180% of regional headroom against a 60% grid-load ceiling, and only 9 ha of measured open land against 661 ha needed. None of the searched alternatives was shown to be viable, so changing the supply mix alone does not rescue this scale at this site. At over 12 years to full build-out, approval on these terms would lock in ratepayer, land and carbon exposure long before the returns could be realised.

Conditions

Reduce grid draw to within the 60% grid-load ceiling and regional headroom, rather than the proposed 1,180% of headroom, with firm contracted supply before construction.
Raise the environment score from 30 to at least the 45 floor, cutting emissions materially below 11 Mt per year through a lower-carbon supply mix and verified procurement.
Assemble and permit the full land requirement beyond the measured 9 ha of open land, toward the 661 ha the facility and its supply need.
Re-prove economics at the remediated scale and supply mix with at least an 8% IRR and payback within seven years, without relying on rate impacts for other customers.

Risks

Grid interconnection and ratepayer backlash could delay, curtail or politically block a draw that is far beyond regional headroom.
Land assembly beyond the 9 ha measured on site may fail, fragment, or inflate costs across the 661 ha requirement.
The 320 gram per kilowatt-hour grid intensity drives about 11 Mt of annual emissions and leaves the environment score well below its floor.
The 7% annual decay in GPU-hour pricing, a year-five GPU refresh, and 69% utilisation could compress the modelled returns over the long build.
Local opposition could intensify despite a community score of 70, given a negative sentiment shift and the scale, noise, traffic and water demands in the densest data centre market in the world.

Scores

68economic

30environment

70community

For the executive

Do not proceed with a 4,000 MW, all-grid build in Loudoun County, even though the headline case shows a 12% IRR, payback in under four years and about $18B in net present value. The binding constraints are physical and political — grid headroom, land assembly and environment — so redirect effort to a smaller, firm-powered design or another site before committing the roughly $172B capital programme.

For the policy maker

This proposal would draw 1,180% of regional grid headroom and emit about 11 Mt of carbon per year, while creating about 3,600 permanent jobs, roughly $2B a year in tax revenue and about $1B a year in local spend. Approval should be withheld until the operator protects other ratepayers, secures land without displacement pressure, and brings environmental performance above the required floor.

For the investor

The proposed returns clear the return floors, but they are not investable as structured because the project fails environment and grid-load floors and depends on power and land that are not secured. Underwrite nothing until firm supply, land control, and a compliant emissions profile are contracted, as delay, curtailment and rate intervention could erase the modelled $50B in annual EBITDA.

How to build it so the verdict holds

This plan does not authorise the NO_GO proposal as drawn; it restates the draft 4,000 MW B200 build only as a gated, phased programme that must stop unless land, firm supply and environmental compliance are secured first. It retains the draft's half gas and half solar supply concept, a peak crew of 2,500 with 20% automated construction, 80 kW direct-liquid density, and 32 staggered halls, each energised only when its supply and community commitments are in place.

Play this plan in the simulation

Configuration

Supply mix	50% on-site · 50% solar
IT load	4,000 MW
Hardware	B200 · 80 kW racks
Crew · automation	2,500 · 20%
Plan scores	63 · E 56 · Env 64 · C 72
Live in · payback	149 mo · 4.2 yrs
Clears the floors	no

Phasing

Gate 0 — no build: no land close or construction until full land assembly, community agreement, permits and firm supply for the phase are secured, reflecting that measured open land is only 9 ha against 661 ha needed.
Gate 1 — enabling works: permitting, community agreement and supply procurement proceed in parallel, but civil works start only when the grid-load and environment floors can be met for the capacity to be built.
Phase 1: first 148 MW hall with B200 racks at 80 kW density, energised only as contracted supply lands and commissioning confirms PUE and noise and traffic compliance.
Phases 2 to 32: each subsequent 148 MW hall is gated individually on secured land, landed firm supply, community-score monitoring above the 40 floor, and re-proven economics, with a stop right if any gate fails.

Community commitments

Local-first hiring for construction trades, tied to the 2,000 construction jobs and about 3,600 permanent jobs, with published local employment reporting through a community liaison.
Published and enforceable noise and traffic limits, with monitoring against the proposal's noise and traffic indices and a funded complaint and remediation process overseen by the liaison.
Heat-reuse or water-offset agreement with the municipality, with annual public reporting on water use, local spend and tax contributions.
Ratepayer-protection commitment that no phase energises until its share of firm supply is contracted, so other customers are not exposed to grid shortfalls.

Energy mixes the advisor searched

Mix	Viable	Score	Payback	IRR	CO₂ kt/yr
Gas + solar 50% on-site · 50% solar	no	63	4.2 yrs	7%	7,249
Renewables + SMR 50% small · 30% solar · 20% wind	no	63	4.6 yrs	6%	486
Hydro PPA + grid 60% hydro · 40% grid	no	59	3.6 yrs	12%	4,820
SMR long-term 70% small · 30% grid	no	58	3.8 yrs	10%	3,612
As proposed (100% grid) 100% grid	no	55	3.5 yrs	12%	11,188
Solar + wind + grid backup 40% solar · 30% grid · 30% wind	no	55	4.7 yrs	5%	3,463
Grid + solar 60% grid · 40% solar	no	54	4.1 yrs	8%	6,690
Grid + wind 60% grid · 40% wind	no	54	4.1 yrs	8%	6,482
Gas bridge 70% on-site · 30% grid	no	53	3.4 yrs	13%	13,595

Energy what-if

Energy supply mix

Grid100%On-site gas0%Solar + storage0%Wind + storage0%Small modular reactor0%Geothermalno mapped resource0%Hydro PPA0%

Sliders are weights; shares are normalized to 100%.

IT load (MW)GPU classRack densityConstruction crewAutomation share 20%

Live estimate

vs. baseline

Composite

Economic

Environment

Community

Metric	This mix	Baseline
Build cost	$172.1B	$172.1B
Months to live	149	149
Blended power	$0.082/kWh	$0.082/kWh
Utilization	69%	69%
EBITDA / yr	$49.6B	$49.6B
Payback	3.5 yrs	3.5 yrs
IRR	12.4%	12.4%
CO₂ / yr	11,188,222 t	11,188,222 t
Carbon intensity	320 g/kWh	320 g/kWh
Water / yr	83,957,746 m³	83,957,746 m³
Grid draw vs headroom	1180%	1180%
Land	661 ha	661 ha
Permanent jobs	3,635	3,635
Nuisance index	50	50

Measured open land is 9 ha; the facility and its supply need 661 ha, so land must be assembled beyond the site (sam-3.1).
Grid draw is 1180% of regional headroom; rate impacts for other customers are likely.

Region profile country priors

Industrial power	$0.082/kWh
Grid carbon	320 g/kWh
Interconnect lead	30 months
Grid headroom	400 MW
Grid reliability	96%
Construction trades	15,000 @ $48/h
Land index	3.20×
Sentiment	38 / 100
Climate · water	temperate · low stress
Solar · wind	0.60 · 0.40
Geothermal · flare gas	no · yes
People within 5 km	unknown

sam-3.1 segmentation of Esri World Imagery

Baseline build cost

Site measured by sam-3.1 over a 1.4 km view: 9 ha open land, 1 ha built, 2 ha water 566 m away, 0 ha solar · Built as 32 phases of 148 MW, staggered 4 months · PUE 1.18 from 80 kW · direct liquid in a temperate climate · GPU-hour rate $3.20 decaying 7%/yr per generation; utilization 69% from supply firmness and fill · Blended power $0.082/kWh, 320 gCO2/kWh · Ten-year cash flow at 10% discount with a GPU refresh in year 5

Scenarios

Name	Mix	Score	Payback	Verdict
Baseline as reportedbaseline	100% grid	55	3.5 yrs	No go

Loading map…