Beijing to build space data center 700-800 km above Earth

Beijing plans a multi-phase space data center in a dawn‑dusk orbit 700–800 km above Earth, designed as a gigawatt‑class centralized system with space computing, relay transmission and ground control subsystems and server clusters totaling millions of cards; a large-scale facility is targeted by 2035. An innovation consortium led by the Beijing Astro‑future Institute and 24 industry partners will integrate the project with AI, mobile communications, new materials and new energy to build new industrial chains; the first experimental satellite is complete and slated for launch in late 2025/early 2026. Government support positions the program as a strategic commercial‑space and AI initiative that could create long‑term opportunities for aerospace, semiconductor and communications suppliers.

Analysis

Market structure: Beijing’s gigawatt-scale space data center is a state-directed demand shock that favors vertically integrated Chinese aerospace contractors, domestic foundries and telecom backhaul operators. Expect improved pricing power for suppliers of radiation-hardened servers, power/solar subsystems and ground-relay equipment through multi-year procurement contracts (CAPEX running into low billions USD annually by 2030). Western cloud hyperscalers (AMZN, MSFT) face competitive displacement for low-latency national workloads; export-controls on high-end GPUs will further tilt hardware spend to domestic suppliers.

Risk assessment: Tail risks include international sanctions escalation, catastrophic on-orbit failure or debris incidents, and slower-than-expected domestic chip progress; any of these could wipe out multi-year revenue assumptions. Timeframes: immediate market noise on announcements (days–weeks), first technical validation hinge 2025–26 launch, full commercialization by 2035; hidden dependencies are domestic high-performance GPU/RF supply and in-space power scaling (>GW is ambitious technologically).

Trade implications: Primary tradeable exposures are Chinese semiconductor (SMIC 0981.HK), telecom infra (China Mobile 0941.HK), and solar/power component makers (JKS) — allocate small starter positions now and scale on concrete procurement signals or a successful 2025 launch. Use option structures (12–36 month calls or call spreads) to express upside while capping downside; consider relative-value pairs long domestic suppliers vs short non-China-exposed premium Western exporters if export-controls intensify.

Contrarian angles: Consensus underestimates technical/thermal/power feasibility and the timescale risk — the 2035 target implies a decade of engineering that can compress returns and raise capital burn. Mispricing opportunity: early-stage HK/A-share suppliers tied to state consortia may be shallowly priced today; avoid binary singles (one supplier) and prefer diversified baskets or ETFs to capture the program while hedging execution risk.

Key Decisions for Investors

• Establish a 1.5–2.5% long position in SMIC (0981.HK) as a thematic play on domestic high-performance logic and packaging for space computing; scale to 4–5% within 12–36 months if Beijing issues formal procurement contracts or SMIC reports >20% YoY revenue guidance tied to aerospace customers.
• Initiate a 1–2% long position in China Mobile (0941.HK) to capture ground-relay and backhaul contracts; add another 1% if the consortium publishes multi-year purchase orders within 6–12 months. Trim to zero if dividend yield falls >200 bps or ARPU declines >5% sequentially.
• Buy 0.5–1% exposure to JinkoSolar (JKS) for terrestrial/space solar tech demand (via stock or 12–24 month call spread). Increase allocation by +1% following a successful 2025/26 satellite power demonstration or if polysilicon contract prices rise >15% signaling sustained demand.
• Execute a pair trade: long 1% SMCI (SMCI) or a hardware-focused basket (12–24 month call spread) vs short 0.5% NVDA (NVDA) via 18–24 month 20% OTM puts as insurance — this expresses a view that localization will steal hardware share in China if export controls persist. Close or invert positions if regulatory headlines reverse within 90 days.