GF Securities released a research report stating that active deployment efforts by China and the US, combined with the inherent cost and performance advantages of space computing power, indicate vast long-term market potential for the industry. As long-term space computing power drives the expansion of solar array power and area, coupled with a surge in single-satellite power requirements and the constraints of launch space and weight, solar arrays will adopt a flexible technology pathway. According to the CGES public account, flexible solar arrays are expected to achieve a 20%-40% weight reduction, shrink the stowed volume by over 60%, and deliver higher performance. Given that solar arrays account for 60%-80% of the value within the power system, this shift in technology is anticipated to make them the most optimal inflationary segment across the entire value chain. The main viewpoints of GF Securities are as follows:
Space computing power is transitioning from a "space data, ground computing" model to a "space data, space computing" paradigm. Essentially, space computing power involves deploying data centers and computational capabilities into space orbit, moving beyond the traditional satellite network model of sending data to Earth for processing to directly performing data computation and processing in space itself.
Space computing power possesses significant advantages in operational costs. While it offers benefits in both performance and operational expenses, the cost advantage is particularly critical. Operational costs are primarily composed of four factors: marginal energy, cooling, water usage, and backup power, with marginal energy being the core component. According to data from LumenOrbit, for instance, the total cost for a single 40MW computing cluster operating in space for 10 years is only $8.2 million, representing potential savings of approximately $159 million compared to a traditional terrestrial computing cluster, with marginal energy costs alone accounting for over $130 million in savings.
Active deployment by China and the US in space computing power points to a vast long-term market space. Based on public accounts such as CCTV Finance, overseas, Nvidia has completed in-orbit testing of its H100 GPU, SpaceX plans to deploy 100GW of space data center capacity annually within four years, Google intends to launch a TPU prototype satellite in 2027, and Starcloud is advancing a 40MW data center project. Domestically in China, a "Zero-Carbon Space Computing Center" has already achieved computing power of 744 TOPS per satellite and 5 POPS for a constellation, with the "Star Computing Plan" aiming to deploy 2,800 satellites to build a global space-based network. The active布局 by both China and the US, layered on top of the inherent cost and performance benefits of space computing, suggests a very expansive long-term market outlook for the industry.
Launch efficiency, space photovoltaics, and performance demands are propelling solar arrays to potentially become the optimal inflationary segment. As long-term space computing power drives the expansion of solar array power and area, coupled with a surge in single-satellite power requirements and the constraints of launch space and weight, solar arrays will adopt a flexible technology pathway. According to the CGES public account, flexible solar arrays are expected to achieve a 20%-40% weight reduction, shrink the stowed volume by over 60%, and deliver higher performance. Current flexible solar array technology pathways are diverging: silicon-based relies on the availability of heavy-lift rockets for widespread adoption, gallium arsenide offers superior performance but at a higher cost, and perovskite represents a promising future route (costing only 1/10th that of gallium arsenide). Since solar arrays constitute 60%-80% of the value within the power system, this technological shift is expected to position them as the most optimal inflationary segment across the entire chain.
Investment recommendation: It is advised to focus on targets related to space photovoltaics. (1) Maxwell Technologies (300751.SZ): According to its official website, the company's main products include fully automatic solar cell screen printing production lines and heterojunction (HJT) high-efficiency cell manufacturing turnkey solutions. It is well-positioned to rapidly enter the aerospace-grade photovoltaic production line market, potentially becoming a core equipment supplier for the photovoltaic segment of space computing power, with significant potential for earnings volatility. (2) Gaoce Technology (688556.SH): Per the company's website, its diamond wire cutting technology is suitable for the thinning of silicon-based solar cells, aligning well with the cost-reduction pathway for space photovoltaics. If space computing power drives a surge in photovoltaic demand, the company's cutting business could open a second growth curve. (3) Jingwei Hi-Tech (300724.SZ): According to its corporate website, the company has a comprehensive layout in TOPCon, HJT, XBC, and perovskite solar cell equipment, providing high-efficiency solar cell equipment and smart manufacturing system solutions. It has the potential to form deep partnerships with downstream domestic aerospace clients, allowing it to benefit from the expansion红利 of flexible solar arrays driven by space computing power.
Risk warnings include intensifying competition, risks associated with technology development and migration falling short of expectations, and risks related to the commercial aerospace sector developing slower than anticipated.
Comments