Orbital Compute Cost of Capital

What is the weighted average cost of capital (WACC) for orbital AI compute infrastructure?

What is the weighted average cost of capital (WACC) for orbital AI compute infrastructure?

Answer

Orbital compute WACC ranges from 10% (optimistic) to 20% (conservative) in 2026, with a central estimate of 13.5%, declining over time as operational history accumulates. By 2040, the central estimate falls to 10%, optimistic to 8%, and conservative to 15%. No orbital compute constellation has been financed, so these estimates derive from satellite constellation financing history, SpaceX's corporate profile, and novel infrastructure analogies.

The central estimate of 13.5% (rather than a naive 15%) reflects a structural adjustment: ~1.5 percentage points of the raw WACC spread duplicate risks already captured in the effective lifetime parameter (see "Double-Counting Adjustment" below). The resulting 6.5pp spread over terrestrial WACC (13.5% vs 7%) reflects technology novelty, revenue uncertainty, limited financing precedent, and the residual volatility of asset-life and catastrophic-loss risks not fully captured in cash-flow modeling.

The historical track record is sobering: of the major LEO constellation ventures, Iridium, OneWeb, ORBCOMM, Globalstar, and Teledesic all went bankrupt or shut down starlink-history.5. Only Starlink avoided this, and only because SpaceX's rocket business provided continuous equity access starlink-history.4.

Analysis

Established Satellite Operator WACCs (Floor, Not Target)

Mature satellite operators demonstrate the floor:

• Iridium: 7.45% WACC, cost of equity 8.72% (beta 0.60–0.71), cost of debt 5.34% iridium-wacc-2025.1
• Eutelsat: 6.23% WACC, cost of equity 6.53%, cost of debt 5.13% eutelsat-wacc-2025.1

These operators have decades of operational history, established revenue, and diversified customers. An orbital compute venture cannot access these rates because it has no operational revenue track record, novel technology with unproven reliability, much shorter asset life (3–5 yr vs 15 yr for GEO), and no secondary market.

At the other end, Viasat's debt/equity ratio reached ~1.5 with 9.0% Senior Secured Notes (Inmarsat legacy, $1.975B) and ~$7B total debt viasat-fy25-financials.1 — demonstrating how quickly satellite company borrowing costs rise with leverage and risk.

Constellation Project Finance (10–16%)

Every successful constellation has required substantial government backing:

• Iridium NEXT: $1.8B Coface-backed credit facility ($1.54B fixed at 4.96%, $0.26B at LIBOR + 1.95%), later refinanced commercially at LIBOR + 3.75% (2019), then SOFR + 2.25% (2023) iridium-next-financing.1. The ECA-backed structure enabled below-market rates; without it, rates would have been substantially higher.

• Telesat Lightspeed: C$2.54B in government-backed financing at CORRA + 4.75% (floating, 15-year). Government received ~10% equity warrants telesat-lightspeed-financing.1. The concessional rate implies commercial borrowing would have been substantially higher.
• OneWeb: went bankrupt during COVID-2020, later acquired by Eutelsat starlink-history.2. Post-merger, the French state backed ~€975M in export credit financing for 440 replacement satellites, alongside a €1.5B government-led equity raise eutelsat-oneweb-eca-financing.1 — another example of government support enabling constellation financing at below-market rates.

Without government backing, constellation project finance rates of 12–16% are implied by the substantial concessions embedded in these government-backed structures.

SpaceX Corporate (8–12%)

SpaceX's financial profile is unique among space companies: ~$800B valuation (Dec 2025), $15.5B revenue, $10B Starlink revenue spacex-financial-profile.1. SpaceX raised $9.1B since 2015, including $1.9B during COVID-2020 (when OneWeb went bankrupt starlink-history.4) — demonstrating ability to access capital in all market conditions. SpaceX also acquired xAI and agreed to buy EchoStar spectrum for $17B spacex-financial-profile.2.

If SpaceX finances orbital compute as a Starlink extension, leveraging existing infrastructure and revenue, the effective WACC could be 8–12% — similar to a mature infrastructure company. However, orbital compute would carry a premium even on SpaceX's balance sheet due to technology novelty and unproven revenue.

Analogous Novel Infrastructure

• Offshore wind: started at 10–12% baseline WACC, driven down to 6–7% over decades with blended finance and government support. Emerging markets: Philippines 11.72% reduced to 6.54% with concessional financing offshore-wind-wacc.1. Key lesson: technology maturation takes decades to compress WACC.
• Subsea cables: tech giants increasingly self-finance major cable builds, with project-level returns likely in the 8–15% range depending on ownership structure. No published WACC for subsea cable projects was found; this is an order-of-magnitude estimate based on infrastructure project finance norms.
• Industry benchmarks: Technology sector 8.5–12%, Energy & Natural Resources 9.0–13.5% industry-wacc-benchmarks-2025.1.

Double-Counting Adjustment

The effective lifetime parameter already captures the expected cash-flow impact of shorter asset life, catastrophic satellite failures, and capacity degradation by reducing the number of capacity-years over which capex is amortized. Standard finance practice prices only residual uncertainty (volatility, tail risk, illiquidity) through the discount rate, not expected losses already in the cash flows. The WACC spread decomposition below shows that ~1.5pp of a naive 15% estimate would duplicate risks already in the effective lifetime model. Removing this overlap yields the adjusted central of 13.5%.

WACC Spread Decomposition

The adjusted central 6.5pp spread over terrestrial (13.5% vs 7%) decomposes as:

Items marked "Yes" or "Partial" are risks whose expected cash-flow impact is already captured by the effective lifetime parameter (which shortens the amortization period to reflect failures, degradation, and obsolescence). The overlap is ~1.5pp — primarily the catastrophic-loss and shorter-asset-life premiums that duplicate what effective lifetime already models. The model uses the adjusted central of 13.5% (net of overlap) rather than the naive 15%.

Scenario Assumptions

Optimistic (10% in 2026, declining to 8% by 2040). SpaceX-financed as a Starlink infrastructure extension, leveraging $10B+ annual Starlink revenue, proven satellite manufacturing, and internal launch capability. Still 3–5pp above hyperscaler terrestrial WACC in 2026 due to technology novelty and shorter asset life. By 2040, converges toward mature infrastructure rates as operational history accumulates. This is the scenario implied by SpaceX's Jan 2026 FCC filing and the SpaceX-xAI merger.

Central (13.5% in 2026, declining to 10% by 2040). Orbital compute evaluated on its own economic merits — either as a SpaceX business unit with its own risk-return profile, or an independent venture with SpaceX as launch provider. The 2026 value of 13.5% is comparable to early satellite constellation financing (pre-operational) and early offshore wind in emerging markets (11–12% baseline), adjusted downward by ~1.5pp to remove double-counting with effective lifetime. The decline to 10% by 2040 follows the offshore wind precedent (10–12% → 6–7% over ~20 years offshore-wind-wacc.1), compressed to a 14-year window reflecting the faster pace of technology iteration in space.

Conservative (20% in 2026, declining to 15% by 2040). Standalone venture without SpaceX's advantages, or early deployment before revenue demonstration. Reflects the catastrophic historical failure rate of LEO constellations in 2026. Even in this scenario, some WACC compression is expected by 2040 as the industry accumulates operational data and financing precedents develop. Approaches but does not reach VC-level return expectations (20–30%+ space-vc-returns.1).

Addressed: Double-Counting with Effective Lifetime

The naive WACC spread of ~8pp (15% vs 7%) includes premiums for "shorter asset life" (+1-2pp), "zero salvage value" (+0.5-1pp), and "self-insurance / catastrophic loss risk" (+0.5-1pp). However, the model's effective lifetime parameter already captures the expected cash-flow impact of these risks. In standard finance, risks already modeled in the cash flows should not also appear as a discount rate premium. The overlap is partial, not total — shorter asset life affects the volatility of cash flows (not just expected values), and zero salvage value eliminates downside cushion — but ~1.5pp of the raw spread duplicates effective-lifetime modeling. The model's central WACC of 13.5% reflects this adjustment.

WACC Compression Over Time

The model now uses time-varying WACC, declining linearly from 2026 values to 2040 values, following the offshore wind precedent where WACC compressed from 10–12% to 6–7% over roughly two decades offshore-wind-wacc.1. The central trajectory (13.5% → 10%) reflects the expectation that operational history, demonstrated revenue, and financing precedents will gradually reduce the risk premium — conditional on the technology succeeding. Early failures could instead push WACC higher, which is why the conservative trajectory remains elevated (20% → 15%).

The compression rates by scenario:

• Optimistic: 10% → 8% (2pp over 14 years) — already near infrastructure rates, limited further compression
• Central: 13.5% → 10% (3.5pp over 14 years) — moderate compression, consistent with offshore wind trajectory
• Conservative: 20% → 15% (5pp over 14 years) — some maturation even in pessimistic case, but remains well above terrestrial

Evidence

1. Iridium Communications WACC: 7.45%. Cost of equity 8.72% (beta 0.60–0.71), cost of debt 5.34%. Market cap ~$3.5B, book debt ~$1.7B. Reflects a mature, post-bankruptcy operator with established government contracts. — iridium-wacc-2025

2. Eutelsat Communications WACC: 6.23%. Cost of equity 6.53%, cost of debt 5.13%. Debt weight 49.6%. Mature GEO operator merged with OneWeb's LEO assets. — eutelsat-wacc-2025

3. Viasat debt/equity ratio ~1.5 (FY2025, March 31 2025 balance sheet: ~$7.0B total debt vs ~$4.6B equity). Inmarsat's 9.000% Senior Secured Notes due 2029 ($1.975B principal). Demonstrates high cost of debt for leveraged satellite operators. — viasat-fy25-financials

4. Iridium NEXT financed with $1.8B Coface-backed credit facility: $1.54B tranche at fixed 4.96%, $0.26B tranche at LIBOR + 1.95%. Refinanced Nov 2019 at LIBOR + 3.75% ($1.45B term loan, no ECA backing). Refinanced again Sep 2023 to SOFR + 2.25% (amended Jun 2024). — iridium-next-financing

5. Telesat Lightspeed secured C$2.54B in government-backed financing (C$2.14B federal + C$400M Quebec) at CORRA + 4.75% (floating, 15-year). Government of Canada received warrants for ~10% of Telesat LEO common shares. — telesat-lightspeed-financing

6. OneWeb went bankrupt during COVID-2020 and was later acquired by Eutelsat. — starlink-history

7. Post-merger, Eutelsat secured ~€975M in French state-backed export credit financing (via Bpifrance Assurance Export, Feb 2026) for procurement of 440 replacement OneWeb LEO satellites from Airbus. Combined with a €1.5B shareholder capital raise led by French and UK governments, total financing covers ~€2.2B for constellation replenishment. — eutelsat-oneweb-eca-financing

8. SpaceX reached ~$800B valuation (Dec 2025). Revenue ~$15.5B (2025), Starlink ~$10B (64%). Planning $1.5T+ IPO targeting mid-2026. — spacex-financial-profile

9. SpaceX acquired xAI and agreed to buy EchoStar spectrum for $17B (up to $8.5B cash plus up to $8.5B in SpaceX stock). — spacex-financial-profile

10. SpaceX raised $9.1B since 2015, including $1.9B during COVID-2020 (when OneWeb went bankrupt). Unique ability to raise equity in all market conditions. — starlink-history

11. Historical LEO constellation failure rate: Iridium (1999 bankruptcy, $1.5B default), OneWeb (2020 bankruptcy), ORBCOMM (bankruptcy), Globalstar (bankruptcy), Teledesic (shut down). Only Starlink avoided bankruptcy. — starlink-history

12. Space VCs cite 10–15x ROIC minimum benchmark. Space hardware requires 5–7 years from founding to meaningful revenue. Over $47B in private capital invested in space since 2015. — space-vc-returns

13. Offshore wind baseline WACC 10–12% (Philippines 11.72%, Vietnam 12.23%), driven down to 6–7% with blended finance combining commercial debt, concessional loans, export credit guarantees, and grants. Philippines: 11.72% reduced to 6.54%. — offshore-wind-wacc

14. [Removed — this was an analytical observation, not sourced evidence. See "Analogous Novel Infrastructure" in the Analysis section above, where the subsea cable 8–15% range is noted as a derived estimate.]

15. 2025 industry WACC benchmarks: Technology 8.5–12.0%, Energy & Natural Resources 9.0–13.5%, Real Estate 5.5–8.5%. — industry-wacc-benchmarks-2025

16. Of ~10,000 active satellites, only ~300 carry insurance (mostly GEO). Fewer than 50 LEO satellites insured. LEO operators overwhelmingly self-insure. Launch premiums 5–20% of satellite value. — leo-insurance-market

17. SpaceX acquired xAI and filed an FCC application to launch up to a million satellites. Separately, Musk stated that "launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually." Musk claims this would be "the lowest cost way to generate AI compute." — spacex-xai-merger