Blue Origin joins the orbital data center race

Blue Origin filed with the FCC for Project Sunrise proposing up to 51,600 sun‑synchronous satellites to provide in‑space computing for AI workloads. The filing seeks Ka‑band authorization, relies on optical intersatellite links tied to Blue Origin’s TeraWave broadband constellation and New Glenn launch capacity, requests a waiver from FCC deployment milestones, and commits to deorbit satellites within five years; the plan competes with SpaceX’s 1,000,000‑satellite filing and Starcloud’s 88,000‑satellite proposal, raising regulatory, orbital‑debris and astronomy‑impact considerations.

Analysis

The entry of large-scale orbital compute proposals reframes the battleground from pure launch volume to specialized subsystems and operations: high-efficiency PV arrays, optical intersatellite terminals, radiation-hardened accelerators, and active de-orbit/TT&C services will capture most of the early margin. Expect supplier bottlenecks with component lead times of 18–36 months and price power concentrated in a handful of producers, which creates a multi-year window where prime contractors (and their subcontractors) can earn outsized margins versus commoditized satellite assemblers.

Regulatory and operational risks dominate the near-to-medium term economics. If regulators tighten debris and brightness rules within 12–36 months, per-satellite OpEx/CapEx could rise by 20–30% from mandatory collision-avoidance fuel, active disposal systems, or ground-track limitations — a shift that would materially extend payback periods and lower utilization thresholds needed to reach parity with terrestrial datacenters. Launch reliability and cadence are single points of failure: a 2–4 year slip or higher failure rate raises replacement costs and financing spreads, making the business case highly path-dependent.

Consensus underestimates the segmentation of workloads that will move off Earth and the intermediate market opportunities. Latency-tolerant, batch ML training and archival hot storage are natural early adopters, but true cost arbitrage requires sustained >40% constellation utilization and multi-year duty cycles. That makes component/launch suppliers and Ka-band ground infrastructure better upstream long candidates than a broad, immediate impairment to terrestrial datacenter landlords — the displacement is asymmetric and likely measured in 3–7 years rather than months.

Key Decisions for Investors

• Long L3Harris (LHX) — 18–36 months. Buy LHX equity or 18–24 month call spreads sizeable enough to capture contract awards for intersatellite optics/TT&C. Upside: +25–40% on contract capture; downside: -15% on defense spending softness. R/R ~2:1.
• Long Maxar Technologies (MAXR) — 12–24 months. Buy shares or LEAP calls to play demand for spacecraft buses, integration, and mission ops. Upside: +30% on increased procurement; downside: -25% if competition commoditizes bus builds.
• Pair trade: Long LHX or MAXR / Short Equinix (EQIX) or Digital Realty (DLR) — 24–60 months. Size 1:1 to express a view that orbital compute will gradually displace a subset of hyperscale demand for price-insensitive training workloads. Target asymmetric payoff: +20–35% on longs vs -10–15% on shorts if thesis accelerates; hedges macro datacenter cyclicality.
• Long Viasat (VSAT) — 12–36 months. Buy VSAT or 12–24 month calls to capture Ka-band ground infrastructure and gateway upgrades; risk is incumbent ground-market competition compressing margins.
• Tactical options: Buy Rocket Lab (RKLB) 12-month calls as a hedge against launch bottlenecks. If small/medium launch demand rises from constellation rollout, RKLB re-rates quickly; if New Glenn/others dominate, premium could evaporate—keep position size limited to 1–2% of risk budget.