A new investment race is unfolding in Low Earth Orbit; here’s why Big Tech and governments are piling in

Investment in LEO crossed $45 billion in 2025, up from under $25 billion a year earlier, indicating rapid capital inflows into satellite infrastructure. Major players are scaling aggressively — SpaceX operates >9,500 satellites with proposals up to 1,000,000, Nvidia has launched a space-computing platform, Amazon plans >3,000 satellites plus approvals for 4,500, Blue Origin >5,000 by 2027, Eutelsat/OneWeb ~600 satellites backed by €1.35bn, and China has filed plans for >200,000 — driven by LEO's lower latency and reduced launch costs. The shift represents a multi-decade infrastructure cycle (>$400bn invested since 2009) and potential market catalysts (e.g., a SpaceX IPO), but regulatory frameworks are lagging, creating operational and liability risks for investors.

Analysis

Space-based compute and connectivity convert a one-off satellite CAPEX cycle into a multi-decade, sticky demand stream for high-margin edge compute, custom accelerators and integrated software stacks. That favors firms that can capture platform-level economics (hardware + software + service) rather than pure-play launch or manufacturing vendors; NVDA sits at the intersection of that demand but will need product and warranty adaptations (radiation tolerance, lifetime service contracts) that extend product cycles and raise ASPs.

Amazon’s LEO play behaves like a customer-acquisition and distribution capex program: it can compress customer lifetime costs in emerging markets and create AWS lock-in via bundled connectivity+cloud contracts. Near-term P&L will be capex-heavy and lumpy, but if execution hits, AMZN gains a structural growth leash into underpenetrated markets; incumbents in terrestrial connectivity will either partner (revenue uplift) or see EBITDA compression from lost high-margin transit services.

Regulation, debris risk and spectrum disputes are the dominant path-dependent catalysts — a serious collision or coordinated moratorium could erase 12–36 months of deployment value and spike insurance/financing costs. Market catalysts to watch are (1) major procurement wins for space-graded GPUs or software, (2) regulatory rulings on spectrum/safety, and (3) a sector re-rating triggered by a large-cap IPO; each maps to distinct trading horizons (days for news/volatility, months for procurement wins, years for revenue realization).

Key Decisions for Investors

• NVDA — directional, 12–18 month asymmetric exposure: buy NVDA Jan 2027 LEAP call spread (buy Jan 2027 $600 call, sell Jan 2027 $900 call) sized to 2–3% of tech sleeve. Rationale: capture secular LEO GPU TAM while financing some premium. Risk/reward: ~3x upside if NVDA sustains data-center growth; hard stop if NVDA falls 25% from entry or implied vol collapses >40%.
• AMZN — core overweight with downside protection, 24–36 month horizon: buy AMZN shares and finance a portion of position by selling 6–9 month out-of-the-money calls (covered-call collar), or alternatively buy Jan 2028 $150 calls (25% deltas) for levered upside. Rationale: capture long-term AWS + LEO distribution optionality; downside hedge protects against capex/regulatory shocks. Target: 2x upside over 3 years, limit drawdown to ~20% via collars.
• Event/volatility tactical: buy near-term NVDA straddle or call fly around major product/earnings/GTC events (2–6 weeks) sized 0.5–1% of portfolio to monetize information and sentiment swings tied to LEO compute announcements. Exit on event close; cut if IV increases >50% pre-event without fundamental catalyst.