China's Orbital Chenguang Lines Up $8.4 Billion to Build a Gigawatt Data Center in Sun-Synchronous Orbit
Beijing-backed Orbital Chenguang secured $8.4 billion in credit lines from twelve Chinese banks to build a gigawatt-scale orbital computing constellation by 2035, opening a state-aligned third front against SpaceX and Blue Origin in the race to put AI compute in orbit.
Overview
A Beijing-incubated startup has become the first Chinese company to formally enter the orbital data center race at scale, securing 57.7 billion yuan ($8.4 billion) in strategic credit lines from a dozen major Chinese banks. Beijing Orbital Twilight Technology Co., Ltd., known as Orbital Chenguang, announced the completion of its Pre-A1 equity round on April 20 alongside the credit framework, according to SpaceNews. The company aims to deploy a gigawatt-scale computing constellation in sun-synchronous orbit by 2035, joining SpaceX and Blue Origin in betting that AI workloads will eventually be cheaper to run in space than on the ground.
What We Know
The credit lines come from twelve state-linked banks, including the Bank of China, Agricultural Bank of China, Bank of Communications, Shanghai Pudong Development Bank, and CITIC Bank, SpaceNews reported. The framework agreements represent commitments for future financing rather than disbursed capital, but the scale signals that orbital compute has been elevated to a national priority within China’s space planning.
Orbital Chenguang was incubated by the Beijing Astro-future Institute of Space Technology, which is itself backed by Beijing’s municipal science and technology commission and the Zhongguancun Science Park administration, SpaceNews reported. The institute leads a consortium of 24 organizations spanning the supply chain. Equity investors in the funding round include Haisong Capital, CITIC Construction Investment Capital, Cathay Capital, and InnoAngel Fund.
The technical concept mirrors the architecture being pursued by US competitors. Orbital Chenguang plans to operate satellites in a dawn-dusk sun-synchronous orbit at 700 to 800 kilometers altitude, a configuration that delivers near-continuous solar exposure with minimal eclipses, according to SpaceNews. The vacuum thermal environment is intended to provide passive radiative cooling. The long-term target is a constellation with capacity exceeding one gigawatt, with development unfolding in three phases: core technology and a first computing constellation between 2025 and 2027, integration with Earth-based data centers from 2028 to 2030, and full operational scale-up by 2035.
An experimental satellite, Chenguang-1, was originally slated for launch in late 2025 or early 2026 but does not appear to have flown, SpaceNews reported. A number of undisclosed satellites were lost during the debut flights of the Ceres-2 and Tianlong-3 rockets earlier this year, though SpaceNews did not link those failures directly to Chenguang-1.
Orbital Chenguang is not the only Chinese effort in this space. Zhejiang Lab, a Hangzhou-based research institution, launched the first batch of 12 satellites of its “Three-Body Computing Constellation” on May 14, 2025, with the goal of reaching a 100-satellite scale by 2027, according to Xinhua. State-owned China Aerospace Science and Technology Corporation (CASC) has separately included space-based data centers in its five-year plan, Space.com reported, citing prior CGTN coverage. China’s broader space computing industry is projected to exceed 250 billion yuan, or roughly $36.6 billion, by 2030, according to Xinhua, citing the China Academy of Information and Communications Technology.
A separate Chinese effort, the ADA Space “star compute” plan, targets 2,800 total computing satellites — 2,400 for inference workloads and 400 for model training — with goals of 100,000-petaflop inference capacity and million-petaflop training capacity once the network is fully operational, according to Xinhua. The same report says ADA Space launched a commercial “Prometheus” cloud service platform in March 2026 that it markets as the world’s first space-based computing power cloud platform for enterprises.
Context
The Chinese push lands in the middle of an accelerating Western race to put AI compute in orbit. SpaceX in January filed with the US Federal Communications Commission for permission to deploy up to one million satellites configured as solar-powered data centers, as previously reported. Blue Origin followed in March with an FCC application for 51,600 satellites under a program internally designated Project Sunrise, as previously reported. Y Combinator-backed Starcloud reached unicorn status in March with a $170 million Series A round, as previously reported.
What distinguishes the Chinese effort is the institutional structure. Where SpaceX, Blue Origin, and Starcloud are private companies pursuing FCC authorizations and venture capital, Orbital Chenguang is a state-incubated entity drawing on credit from policy-aligned banks and operating inside a 24-organization consortium. That financing model is closer to how China has historically built strategic infrastructure than to the venture-funded model that has dominated US orbital data center proposals.
What We Don’t Know
The credit lines are framework commitments rather than disbursements, and Chinese banks have flexibility to draw down or withhold capital depending on milestone progress. Neither SpaceNews nor Xinhua disclosed the equity round’s dollar value, the specific drawdown schedule, or how much of the gigawatt target depends on launch costs that have not yet materialized.
The physics of orbital compute remain unproven at any scale meaningful to AI training. Solar power and passive cooling in low Earth orbit are favorable on paper, but the economics of launching, replacing, and inter-linking thousands of satellites across multi-year hardware lifespans have not been demonstrated.
It is also unclear how Orbital Chenguang’s planned constellation will navigate spectrum coordination and orbital congestion in the increasingly crowded 700 to 800 kilometer band, where Starlink, OneWeb successors, and other Chinese megaconstellations already operate or have filed claims. China has filed for 96,714 satellites across two constellations with the International Telecommunication Union, signaling that Beijing intends to establish presence in low Earth orbit at a scale comparable to SpaceX’s Starlink ambitions.
Whether any of the announced orbital data center programs — Chinese or American — can close the gap with terrestrial AI infrastructure on cost, latency, and reliability is the open question the next several years of test launches and demonstration satellites will begin to answer.