Analysts Predict Cheaper AI Computing in Space by 2030

Research group 33FG models orbital solar-powered AI compute architectures and finds that, at current delivery costs (~$2,000/kg) orbital energy costs ~$18–26/W versus ~$12/W in terrestrial data centres, but costs fall rapidly as launch prices drop. Key inflection points: parity with ground at roughly $1,000/kg (or ~$500–600/kg for some HEO Starlink-class designs), ~30% cheaper at $500/kg and ~50% cheaper at $100/kg; reusable Starship with orbital refuelling could make these economics feasible by 2030. Compute-optimized HEO designs show the largest upside if launch costs decline below $500–1,000/kg, while ultra-light Thin-PV remains equipment-cost sensitive; major tech players (Google) and China are already deploying related satellite initiatives.

Analysis

Market structure: Winners will be launch-service providers, satellite OEMs and vertically integrated hyperscalers that can place compute in HEO (Google (GOOGL/GOOG), selective aerospace primes and Maxar-like suppliers); parity is modeled at ~$500/kg and decisive advantage below ~$100/kg (target by 2030). Losers are long-duration, power‑intensive terrestrial data‑centre REITs (EQIX, DLR) and regional utilities exposed to stable DC load growth. Shift in pricing power will favor players that capture both launch and ops (integrated capex) while commoditizing terrestrial energy margins.

Risk assessment: Key tail risks are catastrophic launch failures, rapid tightening of export/space-regulation (ITAR-style or national security bans), and debris/insurance shocks that could reset economics; any of these can wipe out multi-year capex plans. Time windows: negligible market impact in days, active partnering and pilot projects in months (6–24), structural shifts only if launch costs sustainably fall below $500/kg by 2028–2030. Hidden dependency: onboard equipment cost and radiator/thermal tech become dominant once launch costs drop — so cheaper launches alone won’t guarantee wins.

Trade implications: Tactical: overweight GOOGL (2–3% portfolio) for 12–36 months as a strategic builder of orbital solutions, and selective long in MAXR (1–2%) for satellite infrastructure exposure; reduce EQIX/DLR exposure by 20–30% over 6–12 months. Pair trade: long MAXR 1.5% / short EQIX 1.5% as a 3‑year asymmetric bet tied to launch-cost deflation. Options: buy 12–24 month LEAPS calls on GOOGL and MAXR to capture optionality; fund with short 3–6 month covered calls on EQIX to monetize near-term yield.

Contrarian angles: Consensus underestimates equipment/thermal costs — once launch tariffs drop, component scarcity and rad‑hard chip premiums could keep orbital $/W above modelled minima; terrestrial DCs retain advantages (latency, regulation, on‑demand scaling) limiting addressable demand. Historical parallels: earlier infrastructure shifts (offshore manufacturing) took >10 years and created winners on both sides; unintended consequences include insurance spikes and geopolitically driven market bifurcation (U.S. vs China) that could fragment supply chains and create separate winners.

Key Decisions for Investors

• Establish a 2–3% long position in GOOGL (GOOG/GOOGL) over 12–36 months to capture vertically integrated orbital compute initiatives; add if Starship/other launch provider demonstrates sustained sub-$500/kg pricing or Google announces a funded HEO pilot within 12 months.
• Initiate a 1–2% long in Maxar (MAXR) or equivalent satellite OEM exposure and simultaneously reduce exposure to data‑centre REITs (EQIX, DLR) by 20–30% over the next 6–12 months to reflect medium-term risk to DC economics.
• Construct a pair trade: long MAXR 1.5% / short EQIX 1.5% as a 3‑year trade; exit or rebalance if commercial launch cost fails to decline below $1,000/kg by end‑2026 or falls below $500/kg before 2029 (triggering heavier long allocations).
• Use options to express asymmetric views: buy 12–24 month LEAPS calls on GOOGL and MAXR (size 0.5–1% premium each) funded by selling 3–6 month covered calls on EQIX/DLR; this captures upside optionality while generating income against short-term DC resiliency risk.
• Monitor specific catalysts closely for position adjustments: Starship full reuse demonstration, FAA/DOJ regulatory statements, and China’s orbital supercomputer deployments over the next 12–24 months — increase satellite/launch exposure if launch pricing trajectory confirms sub-$500/kg within verified commercial manifests.