Technical Feasibility Summary¶

Ocean-Based sCO2 Cooling for Hyperscale Data Centers¶

Document ID: CCNL-TFS-2025-01 | Version: 1.0 | Date: December 2025

Purpose: Executive summary for engineering review and validation

1. Concept Overview¶

A closed-loop supercritical CO2 (sCO2) cooling system using cold ocean water (2-4°C) from the Labrador Current to reject 100 MW of thermal load from a shore-based data center.

┌─────────────────────┐
│   DATA CENTER       │
│   (100 MW IT load)  │
│         │           │
│    ┌────┴────┐      │
│    │ Primary │      │      SHORE
│    │   HX    │      │      ─────────────────────
└────┴────┬────┴──────┘
          │
    ┌─────┴─────┐              SEABED
    │  16 pipes │              (200-400m depth)
    │  500mm    ├──────────────────────────┐
    │  sCO2     │     5 km out             │
    └───────────┘                    ┌─────┴─────┐
                                     │ Heat      │
                                     │ rejection │
                                     │ to 3°C    │
    ┌───────────┐     5 km return    │ seawater  │
    │  Return   │◄───────────────────┴───────────┘
    └─────┬─────┘
          │
      To pumps

Key differentiator: Servers remain on shore (accessible for GPU upgrades) while benefiting from passive ocean cooling - avoiding the serviceability limitation that caused Microsoft to discontinue Project Natick.

2. Critical Design Parameters¶

3. Heat Transfer Analysis¶

3.1 Why sCO2?¶

3.2 Heat Transfer Calculation¶

Limiting factor: Seawater-side natural convection (ho ≈ 500 W/m²·K)

Solution: External longitudinal fins (8 per pipe, 50mm height) - Surface area enhancement: 3× - Effective U-value: 24 W/m²·K

Required heat transfer area:

LMTD = ((40-3) - (10-3)) / ln(37/7) = 18.0°C

A = Q / (U × LMTD)
A = 100 MW / (24 W/m²·K × 18 K) = 231,500 m²

Provided: 16 × π × 0.5m × 10,000m = 251,000 m² ✓ (8% margin)

3.3 Pressure Drop¶

ΔP = f × (L/D) × (ρv²/2)
ΔP = 0.01 × (10,000/0.4) × (600 × 0.79²/2)
ΔP = 0.47 bar (acceptable)

4. Pipe Design¶

4.1 Construction¶

4.2 Pressure Verification¶

Hoop stress: σ = P × D / (2 × t)
           = 10 MPa × 500mm / (2 × 25mm)
           = 100 MPa

Steel yield (X52): 360 MPa
Safety factor: 3.6 ✓

4.3 Redundancy¶

• 16 parallel pipes allows continued operation if 4 pipes fail
• N+4 redundancy provides 99.9999% availability for pipe system
• Can isolate individual pipes for maintenance

5. Key Technical Risks¶

Critical unknowns requiring validation:¶

1. Fin efficiency in natural convection at seabed conditions
2. Long-term steel/epoxy coating integrity at 200-400m depth
3. sCO2 pump reliability at 85 bar continuous operation
4. Installation feasibility for finned pipe (handling, burial)

6. Cost Summary¶

6.1 Capital Cost¶

Unit cost: $1,270/kW cooling capacity

6.2 Operating Cost¶

6.3 Comparison to Alternatives¶

Ocean sCO2 breaks even at year 6-8, then saves $10-15M annually.

7. Technology Readiness¶

Recommendation: Conduct 1-5 MW pilot to validate system integration before committing to 100 MW build.

8. Comparison to Precedents¶

CCNL advantage: Combines proven elements (shore-based DC, subsea pipes, cold water) in novel configuration that addresses limitations of prior approaches.

9. Validation Requirements¶

Before proceeding to full-scale, validate:

9.1 Desktop Studies¶

• Third-party thermal model verification
• Detailed FEED study by marine engineering firm
• sCO2 pump vendor engagement (Flowserve, Sulzer)

9.2 Physical Testing¶

• Finned pipe heat transfer testing in seawater tank
• sCO2 circulation loop (lab scale, 10-100 kW)
• Coating durability testing (accelerated marine exposure)

9.3 Pilot Project¶

• 1-5 MW demonstration system
• 12-month operational validation
• Performance vs model comparison

10. Conclusion¶

The ocean sCO2 cooling concept is technically feasible based on:

1. Proven components: Subsea pipelines, sCO2 handling, shell & tube HX
2. Favorable physics: 10-40°C sCO2 vs 3°C ocean = large LMTD
3. Economic advantage: 50-100× lower pumping energy than chillers
4. Environmental benefit: Zero water consumption, no thermal plume

Key uncertainties center on: 1. Finned pipe performance and installation 2. System integration at scale 3. Long-term reliability

Recommended next step: Commission third-party engineering review followed by 1-5 MW pilot project.

CCNL-TFS-2025-01 v1.0