Under constructionGoprofile from priorscurated
Hyperion
Meta · Richland Parish, Louisiana, United States · 2,000 MW · online 2028
Multi-gigawatt campus in northeast Louisiana, among the largest single sites announced.
Supply as reported: New gas generation under utility contract; solar additions announced
Build cost
$67.9B
Months to live
70
gated by 13 construction phases
EBITDA / yr
$18.4B
Payback
3.7 yrs
IRR
10.8%
CO₂ / yr
3,982 kt
296 g/kWh
Jobs
2,000 / 1,460
build / permanent
Composite
60
E 64 · Env 46 · C 72
Goon “Baseline as reported” (52% on-site · 30% solar · 17% grid) · Oct 30, 2026 · meta/muse-wonder-1.0
The verdict is GO: Hyperion, Meta's 1,600 MW campus in Richland Parish, Louisiana, clears every floor, with an 11% IRR against an 8% minimum and payback in about four years against a seven-year maximum. It posts a $2.2B ten-year NPV on about $68B of capex, a community score of 72 against a 40 floor and a grid load share of 41% against a 60% ceiling. The case is not comfortable on environment, scoring 46 against a 45 floor with about 4 Mt of CO2 a year, and it needs 1,547 ha when only 10 ha of open land was measured on site. The viable Gas + solar alternative scores higher overall and on environment, but at a lower 9% IRR, so the as-proposed mix remains the economic choice if its conditions are enforced. Proceed in phases, gated by land assembly and firm supply, rather than as an unconditional commitment.
Conditions
- Assemble the full 1,547 ha needed for the facility and its supply, against only 10 ha of measured open land on site, before committing to the full build.
- Hold the environment score above its 45 floor, where the proposal scores 46, through delivery of the on-site gas and solar supply as modelled.
- Keep grid draw within the 60% grid-load-share ceiling, where the proposal uses 41%, and keep blended power near $0.04 per kWh to protect the 11% IRR.
- Secure the 13 phases of supply, racks and GPUs in sequence so utilization, payback and the year-5 GPU refresh remain as modelled.
Risks
- Land is the binding constraint, with 1,547 ha required against 10 ha measured open, so failure to assemble land beyond the site would stop or shrink the campus.
- The environment score of 46 sits only one point above the 45 floor, leaving almost no margin for higher carbon intensity, water use or supply slippage.
- Emissions of about 4 Mt a year and water use of about 19 million cubic metres a year will draw sustained scrutiny even though water stress is not flagged.
- A 70-month, 13-phase build on about $68B of capex exposes returns to GPU pricing, the year-5 refresh and utilization holding near 63%.
- If on-site gas or solar delivery slips, greater grid reliance could push grid load share toward the 60% ceiling and erode power cost and IRR.
Scores
64economic
46environment
72community
For the executive
This is a GO for Meta's 1,600 MW Hyperion campus, but execution is gated first by land assembly and then by firm on-site supply landing ahead of each hall. Prioritize the 52% on-site, 30% solar and 17% grid mix that preserves a 3.7-year payback while keeping the environmental margin defensible.For the policy maker
The campus brings 1,460 permanent jobs, about $861M a year in tax revenue and about $305M a year in local spend, with a community score of 72 against a 40 floor. Approval should lock in published noise, traffic and water commitments and require reporting on 4 Mt a year of carbon emissions before later phases proceed.For the investor
The proposal clears the investment floors with an 11% IRR against an 8% minimum, a $2.2B ten-year NPV and payback in under four years on about $68B of capex. The return cushion is real but thin on environment and land, so funding should be tranched by phase against land close, supply commissioning and utilization.How to build it so the verdict holds
Build the 1,600 MW IT, 1,888 MW facility campus as proposed with B200 GPUs at 80 kW density, in 13 phases of 145 MW staggered about four months apart, reaching full operation in about 70 months. Keep the viable as-proposed supply mix of 52% on-site gas, 30% solar and 17% grid, a 2,500-person peak crew and 20% construction automation, energizing racks only as firm supply lands.
Configuration
| Supply mix | 52% on-site · 30% solar · 17% grid |
| IT load | 1,600 MW |
| Hardware | B200 · 80 kW racks |
| Crew · automation | 2,500 · 20% |
| Plan scores | 60 · E 64 · Env 46 · C 72 |
| Live in · payback | 70 mo · 3.7 yrs |
| Clears the floors | yes |
Phasing
- Permitting, community agreement and assembly of the full 1,547 ha before land close and full commitment
- On-site gas and solar supply procurement in parallel with civil works, gated by firm supply for the first hall
- Phase 1: 145 MW hall, racks energized only as supply lands
- Phases 2 to 6: 145 MW halls staggered about four months apart, each gated by commissioned supply and prior-phase utilization
- Phases 7 to 13: remaining 145 MW halls staggered about four months apart, each gated by supply, land, environmental reporting and the year-5 GPU refresh plan
- Full 1,600 MW operation at about 70 months, gated by final commissioning and community compliance reporting
Community commitments
- Local-first hiring for construction trades, with the 2,500-person peak crew drawn first from the 9,000-person local labor pool and reporting on local hires each phase.
- Published noise and traffic limits with a named community liaison, regular public reporting, and corrective action if limits are breached.
- Heat reuse or water offset agreement with the municipality covering the modelled water use, agreed before Phase 1 energization.
- Sustain the local spend and tax commitments underpinning the 72 community score, keeping it well above the 40 floor, with an annual community impact statement.
Energy mixes the advisor searched
| Mix | Viable | Score | Payback | IRR | CO₂ kt/yr |
|---|---|---|---|---|---|
| Gas + solar 50% on-site · 50% solar | yes | 64 | 4.0 yrs | 9% | 2,916 |
| As proposed (52% on-site · 30% solar · 17% grid) 52% on-site · 30% solar · 17% grid | yes | 60 | 3.7 yrs | 11% | 3,982 |
| Renewables + SMR 50% small · 30% solar · 20% wind | no | 64 | 4.4 yrs | 7% | 195 |
| Hydro PPA + grid 60% hydro · 40% grid | no | 60 | 3.4 yrs | 13% | 2,369 |
| SMR long-term 70% small · 30% grid | no | 60 | 3.7 yrs | 11% | 1,781 |
| Solar + wind + grid backup 40% solar · 30% grid · 30% wind | no | 57 | 4.5 yrs | 6% | 1,689 |
| Grid only 100% grid | no | 55 | 3.3 yrs | 14% | 5,594 |
| Gas bridge 70% on-site · 30% grid | no | 55 | 3.3 yrs | 14% | 5,773 |
| Grid + solar 60% grid · 40% solar | no | 55 | 3.8 yrs | 10% | 3,327 |
| Grid + wind 60% grid · 40% wind | no | 54 | 4.0 yrs | 9% | 3,208 |
Energy what-if
Energy supply mix
Sliders are weights; shares are normalized to 100%.
Live estimate
vs. baselineComposite
60
Economic
64
Environment
46
Community
72
| Metric | This mix | Baseline |
|---|---|---|
| Build cost | $67.9B | $67.9B |
| Months to live | 70 | 70 |
| Blended power | $0.039/kWh | $0.039/kWh |
| Utilization | 62% | 62% |
| EBITDA / yr | $18.4B | $18.4B |
| Payback | 3.7 yrs | 3.7 yrs |
| IRR | 10.8% | 10.8% |
| CO₂ / yr | 3,981,669 t | 3,981,669 t |
| Carbon intensity | 296 g/kWh | 296 g/kWh |
| Water / yr | 19,130,602 m³ | 19,130,602 m³ |
| Grid draw vs headroom | 41% | 41% |
| Land | 1,547 ha | 1,547 ha |
| Permanent jobs | 1,460 | 1,460 |
| Nuisance index | 54 | 54 |
- Measured open land is 10 ha; the facility and its supply need 1547 ha, so land must be assembled beyond the site (sam-3.1).
Region profile country priors
| Industrial power | $0.055/kWh |
| Grid carbon | 400 g/kWh |
| Interconnect lead | 16 months |
| Grid headroom | 800 MW |
| Grid reliability | 96% |
| Construction trades | 9,000 @ $34/h |
| Land index | 0.80× |
| Sentiment | 62 / 100 |
| Climate · water | temperate · low stress |
| Solar · wind | 0.70 · 0.30 |
| Geothermal · flare gas | no · yes |
| People within 5 km | unknown |
Baseline build cost
Site measured by sam-3.1 over a 1.4 km view: 10 ha open land, 0 ha built, 0 ha water, 0 ha solar · Built as 13 phases of 145 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 62% from supply firmness and fill · Blended power $0.039/kWh, 296 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 | 52% on-site · 30% solar · 17% grid | 60 | 3.7 yrs | Go |
Loading map…