Foreword:

On July 26, a conference themed “Accelerating the Strategic Development of Digital Space” in Beijing marked a significant milestone for the commercial space sector—the official launch of the “Digital Space No.1” experimental satellite project. This satellite is no ordinary one; it embodies a profound question: Can satellites develop their own “brain”?

Traditional Models Fall Short: Satellites Must Learn to ‘Operate Autonomously’

For decades, most satellites have followed a rigid operational pattern: orbit-based observation, data transmission to ground stations, processing and analysis by ground systems, and then the relay of instructions back to the satellite for execution. As pointed out by Academician Wei Fengsi from the Chinese Academy of Sciences, by the time this process concludes, “the space environment may have already undergone significant changes.”

Low-Earth orbit satellites operate between 350 kilometers and over 1,000 kilometers above the Earth’s surface, where atmospheric density, ionospheric conditions, and geomagnetic field strength are in constant flux. During solar storms, these parameters can fluctuate wildly within minutes. In February 2022, SpaceX launched 49 Starlink satellites in a single mission, but 40 of them were destroyed upon re-entry into the Earth's atmosphere after encountering a solar storm. By the end of the first quarter of 2026, the overall failure rate of Starlink satellites had reached 12.7%.

Wei Fengsi summarized the core challenges of the new space era as the “Six Highs”: high dynamic variability, high real-time responsiveness, high-precision control, high-resolution identification, high spatial-temporal coverage, and high digital intelligence applications. He bluntly stated that while China’s space industry has made rapid strides in hardware development over the decades, “the soft capabilities of digitization and intelligence have lagged behind.” Most satellites are “equipped with eyes but lack brains.”

Paradigm Shift: From ‘Ground-Based Remote Control’ to ‘Intelligent Satellite Autonomy’

In 2016, Wei Fengsi proposed the “Digital Space” strategic vision to the Chinese Academy of Sciences, advocating for the construction of a space digitalization and intelligence system to meet the challenges of the new space era. In December 2023, the “Digital Space No.1” experimental plan received project support from the Ministry of Science and Technology. By July 2026, the project was officially launched. The mission of “Digital Space No.1” is clear-cut: to comprehensively validate in orbit the “perception-cognition-behavior” brain intelligence architecture proposed by Wei Fengsi.

The “satellite brain” system comprises three subsystems—perception, cognition, and behavior—aiming to enable satellites to autonomously complete the entire process from environmental monitoring to decision-making and execution in orbit, without relying on ground command loops. This mirrors the functions of the human brain—integrating space perception, space cognition, and space behavior into a fully intelligent system onboard the satellite. If China takes the lead in establishing a digital space brain system centered on satellite brains, it could transition from a follower to a leader in the new frontier of intelligent space.

A Dual Approach: From the ‘Huashan’ Large Model to Satellite-Based Laser Communication

The launch of “Digital Space No.1” is not an isolated event; a “space intelligent autonomy” system driven by both hardware and software is taking shape. At the hardware level, satellite-based laser communication is entering mass production. On July 16, 2026, Shaanxi’s first specialized intelligent production line for satellite-based laser communication was inaugurated in Xi’an, capable of producing 600 laser communication terminals annually in its initial phase, with plans to exceed 1,000 units per year by 2027. Jing Zhenlong, Deputy General Manager of CentiSpace, described the alignment precision of laser communication terminals: “It’s like two optical lenses needing to align with each other from thousands of kilometers away and maintain stable, uninterrupted communication for extended periods.” Relying on over two decades of technical accumulation from the Xi’an Institute of Optics and Precision Mechanics at the Chinese Academy of Sciences, CentiSpace has completed the integrated research and development of a new generation of 100 Gbps-class satellite-based laser communication terminal hardware and software.

At the software level, the “Huashan” large model has undergone in-orbit validation. On July 24, 2026, the “Huashan” satellite health management module, an aerospace vertical large model independently developed by CentiSpace, was successfully launched into orbit aboard the “Chenguang No.1” satellite. This model deeply integrates professional knowledge bases across orbital dynamics, optoelectronic payloads, and satellite operations and maintenance. After entering orbit, it continuously iterates and trains using onboard computing power, autonomously monitoring the satellite’s full lifecycle status.

CentiSpace defines this “software-hardware integration” technical system as “space intelligent autonomy”—analogous to autonomous driving for ground vehicles, enabling satellites to achieve fully autonomous in-orbit operation, including perception, transmission, intelligent analysis, orbital avoidance, and fault repair. Satellite-based laser communication terminals establish high-speed inter-satellite data channels, while lightweight large models autonomously perform data analysis onboard, forming a complete closed loop for “space intelligent autonomy.”

Satellite-Ground Synergy: Not Replacing Ground Control, But Redefining Roles

“Intelligent Satellite Autonomy” does not aim to completely sever satellites from ground control but rather to redefine the division of labor between satellites and the ground. Wei Fengsi proposed a new paradigm of “limited autonomy + satellite-ground synergy” in risk scenarios, where routine tasks are completed autonomously by satellites, while complex decisions are made through satellite-ground collaboration.

This logic mirrors ground-based autonomous driving: Level 4 autonomous vehicles do not require human intervention at all times but may still request takeover in scenarios beyond their capabilities. The same applies to “intelligent satellite autonomy,” where satellites autonomously handle perception, cognition, and decision-making in orbit, while the ground provides strategic oversight and intervention. Zeng Weigang, General Manager of CentiSpace, pointed out that compared to traditional manual operations and maintenance, onboard large models can autonomously mine patterns from vast historical data, significantly reducing the frequency of ground control interventions and markedly improving fault identification and emergency response efficiency. This approach is well-suited for the large-scale management and control needs of tens of thousands of satellites operating in orbit simultaneously in the future. Aerospace operations and maintenance are transitioning from a traditional “experience-driven, passive response” model to a “data-driven, proactive defense” intelligent system.

A New Frontier in Commercial Space: 12 Enterprises Join Forces

On the same day as the launch of the “Digital Space No.1” project, a more industrially significant signal was sent: Twelve commercial space enterprises, including Digital Space, Spacety, CentiSpace, Taihu Nebula, Oriental Starlink, Juntian Aerospace, Xingyi Lianxin, and WeChat (assuming WeChat's involvement is metaphorical or represents a tech consortium, as direct space involvement is unlikely), signed industrial cooperation agreements at the conference. The parties will collaborate on experimental satellite development, intelligent payload integration, platform synergy and adaptation, in-orbit testing, and application scenario expansion.

From a broader perspective, “intelligent satellite autonomy” is opening up a brand-new commercial space frontier. Wei Fengsi proposed the technical vision of “intelligent facial recognition”—reconstructing a dynamic, time-varying global space environment image by matching satellite observation data with historical databases, forming global distribution maps of satellite orbits, communication guarantees, and navigation positioning affected by the space environment. If this technology can be implemented, it could serve not only China’s space industry but also global low-Earth orbit constellation operations, forming a new type of infrastructure and industrial chain for commercial space.

Conclusion:

From the proposal of the “Digital Space” strategic vision in 2016 to the official launch of the “Digital Space No.1” project in 2026, the main theme of space development in the next decade will be the digitization and intelligence of space. Only when every satellite possesses its own “brain” will the new space era truly arrive.

Online Sources:

Science and Technology Daily: “Huashan” Large Model Validates Intelligent Satellite Autonomy Technology in Orbit

Economic Observer: “Academician Wei Fengsi of the Chinese Academy of Sciences: Every Satellite Should Have Its Own ‘Brain’”

National Business Daily: “Can Satellite ‘Intelligent Autonomy’ Be Achieved?”