OperationalConditionalprofile from priorscurated
Colossus
xAI · Memphis, Tennessee, United States · 300 MW
Former manufacturing site converted at speed; second phase announced nearby.
Supply as reported: Grid plus on-site gas turbines, battery storage
Build cost
$10.0B
Months to live
30
gated by 2 construction phases
EBITDA / yr
$3.0B
Payback
3.3 yrs
IRR
13.9%
CO₂ / yr
803 kt
384 g/kWh
Jobs
1,152 / 228
build / permanent
Composite
57
E 76 · Env 43 · C 47
Conditionalon “Baseline as reported” (60% on-site · 40% grid) · Oct 30, 2026 · meta/muse-wonder-1.0
The verdict is CONDITIONAL, not a straight go, because the as-proposed 60% on-site gas and 40% grid design is economically strong but breaches the environment floor. It scores 76 on economics with 14% IRR, a 3-year payback and $1,704M ten-year NPV on about $9,961M of capex, yet its environment score is 43 against a 45 floor and emissions are about 803 kt per year. The searched Hydro PPA + grid alternative is viable and better overall, with a composite of 64 versus 57, environment at 60, community at 52, 13% IRR, a 3-year payback and emissions of about 297 kt per year. Community at 47 already clears the 40 floor, grid load share is within its limit, and there is no construction labour shortfall, so the condition is about fixing supply, not abandoning the site. Build it only on the hydro-led mix and the community and phasing commitments in the plan.
Conditions
- Adopt the viable Hydro PPA + grid supply (60% hydro, 40% grid) in place of the as-proposed gas-heavy mix, with the PPA contracted before Phase 1 energization so the environment score rises above the 45 floor.
- Keep community score above the 40 floor through the published noise, traffic, hiring and liaison commitments in the plan, verified before Phase 2 proceeds.
- Hold economics within floors on the hydro mix: IRR at or above 8% and payback at or under 7 years, rechecked at PPA signing.
Risks
- The hydro PPA may not be secured on price or timing assumed, which would push the project back to a mix that fails the environment floor.
- Emissions and disclosure scrutiny remain material if any gas or grid-heavy fallback replaces the hydro supply during delivery.
- GPU economics depend on the assumed $3.20 GPU-hour rate decaying at 7% per year and 69% utilization, so a faster price fall or lower fill would erode the 13% hydro-mix IRR.
- Water use of roughly 3.9M cubic metres per year and local noise and traffic impacts could pull the community score toward its floor if offsets and limits are not enforced.
Scores
76economic
43environment
47community
For the executive
This is a conditional build at Memphis: the 240 MW, B200 design is economically strong, but only the hydro-led supply keeps it approvable. Lock the Hydro PPA + grid mix early and the two-phase, 30-month programme protects speed without carrying the gas mix's environmental breach.For the policy maker
The as-proposed mix fails the 45 environment floor at 43 despite a community score of 47, above its 40 floor, so approval should be tied to the cleaner supply. The hydro mix lifts environment to 60 and community to 52 while the site still delivers 228 permanent jobs and $137M in annual tax revenue.For the investor
As proposed the project shows 14% IRR, 3-year payback and $1,704M NPV over ten years, while the viable hydro mix still returns 13% IRR and a 3-year payback. The return give-up is small against cutting emissions from about 803 kt per year to about 297 kt per year and clearing the environment floor.How to build it so the verdict holds
Build the 240 MW IT (283 MW facility) Colossus expansion in Memphis in two 142 MW phases over about 30 months, using B200 GPUs at 80 kW density with the viable 60% hydro and 40% grid mix instead of the as-proposed gas mix. Use a peak crew of 1,440 with 20% construction automation, energizing each hall only as contracted supply lands so the environment and community floors hold.
Configuration
| Supply mix | 60% hydro · 40% grid |
| IT load | 240 MW |
| Hardware | B200 · 80 kW racks |
| Crew · automation | 1,440 · 20% |
| Plan scores | 64 · E 74 · Env 60 · C 52 |
| Live in · payback | 30 mo · 3.4 yrs |
| Clears the floors | yes |
Phasing
- Permitting and community agreement before land close, including published noise and traffic limits and the community liaison.
- Hydro PPA and grid interconnection procurement in parallel with civil works, gated on signed supply before major power equipment release.
- Phase 1: 142 MW hall with 80 kW direct-liquid racks, energized only as hydro and grid supply lands.
- Phase 2: 142 MW hall staggered after Phase 1, gated on Phase 1 supply performance, environment compliance and the community review.
Community commitments
- Local-first hiring for construction trades, reporting local hires and the $53M in annual local spend to the municipality.
- Heat reuse or water offset agreement with the municipality, addressing the roughly 3.9M cubic metres of annual water use.
- Published noise and traffic limits with a named community liaison, regular public reporting, and corrective action triggers to keep the community score above 40.
- Phase-gate community review before Phase 2 energization, confirming noise, traffic and hiring commitments are being met.
Energy mixes the advisor searched
| Mix | Viable | Score | Payback | IRR | CO₂ kt/yr |
|---|---|---|---|---|---|
| Hydro PPA + grid 60% hydro · 40% grid | yes | 64 | 3.4 yrs | 13% | 297 |
| SMR long-term 70% small · 30% grid | yes | 61 | 3.7 yrs | 11% | 223 |
| Gas + solar 50% on-site · 50% solar | yes | 58 | 4.0 yrs | 8% | 436 |
| Grid + wind 60% grid · 40% wind | yes | 57 | 4.0 yrs | 9% | 399 |
| Grid + solar 60% grid · 40% solar | yes | 55 | 3.9 yrs | 9% | 414 |
| Renewables + SMR 50% small · 30% solar · 20% wind | no | 58 | 4.4 yrs | 6% | 29 |
| As proposed (60% on-site · 40% grid) 60% on-site · 40% grid | no | 57 | 3.3 yrs | 14% | 803 |
| Gas bridge 70% on-site · 30% grid | no | 57 | 3.3 yrs | 14% | 822 |
| Solar + wind + grid backup 40% solar · 30% grid · 30% wind | no | 57 | 4.5 yrs | 5% | 213 |
| Grid only 100% grid | no | 56 | 3.3 yrs | 14% | 692 |
Energy what-if
Energy supply mix
Sliders are weights; shares are normalized to 100%.
Live estimate
vs. baselineComposite
57
Economic
76
Environment
43
Community
47
| Metric | This mix | Baseline |
|---|---|---|
| Build cost | $10.0B | $10.0B |
| Months to live | 30 | 30 |
| Blended power | $0.053/kWh | $0.053/kWh |
| Utilization | 69% | 69% |
| EBITDA / yr | $3.0B | $3.0B |
| Payback | 3.3 yrs | 3.3 yrs |
| IRR | 13.9% | 13.9% |
| CO₂ / yr | 803,494 t | 803,494 t |
| Carbon intensity | 384 g/kWh | 384 g/kWh |
| Water / yr | 3,897,906 m³ | 3,897,906 m³ |
| Grid draw vs headroom | 23% | 23% |
| Land | 45 ha | 45 ha |
| Permanent jobs | 228 | 228 |
| Nuisance index | 32 | 32 |
Region profile country priors
| Industrial power | $0.060/kWh |
| Grid carbon | 330 g/kWh |
| Interconnect lead | 20 months |
| Grid headroom | 500 MW |
| Grid reliability | 96% |
| Construction trades | 9,000 @ $36/h |
| Land index | 1.10× |
| Sentiment | 57 / 100 |
| Climate · water | temperate · low stress |
| Solar · wind | 0.65 · 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: 107 ha open land, 1 ha built, 0 ha water, 0 ha solar · Built as 2 phases of 142 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 $0.053/kWh, 384 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 | 60% on-site · 40% grid | 57 | 3.3 yrs | Conditional |
Loading map…