Jeff Bezos’ Blue Origin enters the space data center game

Blue Origin filed with the FCC on March 19 to launch more than 50,000 satellites under “Project Sunrise” to perform in-orbit computation, using its proposed Terawave constellation as a high-throughput communications backbone. Competing proposals include SpaceX’s 1,000,000‑satellite filing, Starcloud’s 60,000 craft, and Google’s Project Suncatcher, but projects face major near-term hurdles — cooling, laser inter-satellite comms, radiation effects on chips, launch cost sensitivity (market hopes pinned on cheaper Starship flights), orbital congestion, and potential upper-atmosphere/ozone impacts. The filing lacked timing details and experts expect such concepts are unlikely to materialize before the 2030s.

Analysis

Moving substantial compute into LEO is not a single-technology bet but a four-factor economic arbitrage: launch $/kg, radiation-tolerant silicon yield, inter-satellite optical throughput/latency, and thermal rejection mass. Practically, the business case flips from niche to scalable only if launch costs compress ~5-10x from today and if photonic links plus on-orbit cooling push SWaP (size, weight, and power) efficiencies to within ~30-40% of terrestrial rack economics; absent both, projects remain demonstrators with strategic PR value.

Second-order winners are likely component-level: suppliers of space-grade photonics (laser comm terminals), GaN RF/photonic ICs, and lightweight radiator/heat-pipe manufacturers will see concentrated demand early — these companies can earn 3-5x margin expansion from design wins even if full constellations take a decade. Conversely, terrestrial hyperscalers and data-center REITs face a long-tail competitive risk to their cooling/water-intense footprints, but meaningful capex repricing or tenant loss is multi-year and contingent on regulatory/insurance outcomes.

The dominant non-market risks are regulatory (orbital-capacity limits, spectrum allocation, environmental rules tied to upper-atmosphere chemistry) and insurance/collision externalities that could impose per-satellite operating taxes or escrowed de-orbit funds; either would materially change NPV through recurring OPEX increases, not one-time capex. Near-term catalysts to watch are (1) a repeatable low-cost reusability cadence from launch providers, (2) radiation-performance benchmarks for advanced AI accelerators in LEO, and (3) adjusted orbital-debris insurance pricing — any of which can compress the timeline from “2030s” to early 2030s.

Key Decisions for Investors

• Long GOOG/GOOGL LEAPs (buy Jan-2028 $150–$220 call spreads, net debit): 12–36 month horizon. Rationale: captures optionality of winning cloud + space-compute integrations; position size 1.5% notional. Risk: premium decay and execution risk if Google deprioritizes space projects; reward asymmetry 3:1 if Google captures early enterprise partnerships.
• Tactical long PL (Planet Labs) equity or 6–12 month OTM calls (buy Aug-2026 $X calls; size 0.5–1% notional): expect binary re-rate around demo success/partnership announcements. Risk high — demo failure could cut >40%; reward is >2x on successful technology/contract validation.
• Pair trade: long GOOGL LEAPs vs short AMZN LEAPs (equal notional, 12–36 months): expresses view that Google’s cloud/AI positioning benefits more from spatial compute optionality than Amazon (whose founder’s personal ventures create headline risk but limited Amazon corporate arbitrage). Target relative return 2–3x with downside capped to premium paid; rebalance on major regulatory developments.
• Capital allocation hedge: maintain a 1–2% portfolio allocation to suppliers of space photonics/thermal tech via focused small-cap exposure or thematic ETF; trim if orbital-insurance premiums rise >25% or if FCC/ICAO set hard caps on new LEO deployments within 24 months.