AI is reshaping aerospace networks

On June 10th, the Ministry of Industry and Information Technology released the "Implementation Opinions on the Innovative Development of 'Artificial Intelligence + Information and Communication' (2026 - 2028)".

This is an action plan for the information and communication industry. However, when viewed in the context of industrial evolution, its impact will be broader. Artificial intelligence is changing the construction logic of network infrastructure.

With the advancement of applications such as large models, intelligent agents, low - altitude economy, and satellite Internet, the network needs to undertake more functions, including computing power coordination, edge reasoning, intelligent operation and maintenance, real - time perception, and security protection.

For the aerospace industry, this document is also worthy of attention.

The document mentions 5G - A/6G, optical networks, intelligent computing networks, network intelligent agents, and space - based computing networks multiple times. The requirements of artificial intelligence for network capabilities are extending from ground infrastructure to the aerospace system.

I. The Network Shifts from "Connection" to Scheduling

After the implementation of AI applications, the role of communication networks is changing.

In the past, when evaluating a network, the main factors considered were coverage, bandwidth capacity, the number of connections, and transmission stability. Now, new indicators are becoming important.

Large - model training requires cross - regional computing power coordination, and the operation of agents requires continuous data and tool calls. Scenarios such as the low - altitude economy and industrial control put forward higher requirements for latency and edge computing.

The network cannot only be responsible for sending out data; it also needs to participate in computing power scheduling, task distribution, path optimization, and service guarantee.

The document of the Ministry of Industry and Information Technology proposes that by 2028, the information and communication network will initially achieve a high - level of self - intelligence, form more than 30 high - value typical scenarios, and the coverage rate of the 1 - millisecond latency circle of metropolitan computing power will be no less than 75%.

These indicators reflect the changes in network evaluation criteria. Computing power accessibility, end - to - end latency, network self - intelligence ability, and resource utilization are becoming new basic indicators.

This document will have an impact on the aerospace industry. Satellite Internet, low - altitude communication, remote - sensing data transmission, and space - based computing networks all rely on stronger network scheduling capabilities.

When AI applications move from the cloud to the edge and from the ground to the aerospace, network upgrading will become a prerequisite for the continued development of aerospace infrastructure.

II. The Space - Based Computing Network Deserves Special Attention

The "space - based computing network" appears in the part of key core technology research, in parallel with the directions of 5G - A/6G, the new - generation optical network, network - native intelligence, and the intelligent agent Internet.

This shows that the discussion of the future network system is not limited to the ground communication network, and the integration of heaven and earth is entering a more specific technical framework.

In the past, satellites mainly undertook coverage and transmission tasks, helping to solve the connection problems in areas difficult to reach by ground networks. After the acceleration of the construction of low - orbit constellations, the role of the satellite Internet is shifting from a supplementary network to a part of the new - generation information and communication infrastructure.

We can observe this change from several aspects.

First, the network functions undertaken by low - orbit satellites will be more diverse.

In the future, in addition to providing communication services, low - orbit constellations also need to participate in data relay, task distribution, edge computing, and global - scale network coordination. In scenarios such as the ocean, low - altitude areas, emergency support, and remote areas, the role of the space - based network will be further enhanced.

Second, constellation operation will also rely more on intelligent capabilities.

When thousands or even tens of thousands of satellites are in orbit simultaneously, tasks such as inter - satellite link switching, task scheduling, fault handling, and resource optimization are difficult to be completed manually in the long term. Technologies such as network self - intelligence, multi - agent collaboration, and network - native intelligence may enter the constellation operation system in the future.

Another point is that computing power will gradually extend to space. Scenarios such as remote sensing, communication, low - altitude supervision, and emergency rescue will generate a large amount of data. If all the data is transmitted back to the ground for processing, both efficiency and links will be limited.

In the future, a part of data processing work will be completed in orbit, including target recognition, data screening, compression processing, inter - satellite forwarding, and intelligent routing.

Therefore, the focus of the next stage of the satellite Internet is not only on launching more satellites but also on improving the capabilities of the satellites themselves. Computing, perception, routing, and autonomous collaboration will become part of the constellation capabilities.

III. The Intelligent Computing Network Will Drive Space Optical Communication and Inter - Satellite Interconnection

The construction of the intelligent computing network first takes place on the ground.

Large - model training, agent operation, and data center cluster coordination all require a network connection with higher speed, greater stability, and lower latency.

The document proposes that the construction of 400Gbps/800Gbps backbone transmission networks should be accelerated, and research on technologies such as all - optical switching, high - end optoelectronic chips, optoelectronic co - packaged devices, and optoelectronic interconnection of intelligent computing super - nodes should be promoted.

These directions mainly serve the ground intelligent computing centers, but they are not separated from the aerospace industry.

The ground intelligent computing network requires high - speed optical interconnection, and the satellite Internet also requires inter - satellite laser links and high - speed satellite - to - ground transmission.

As the scale of low - orbit constellations expands, data exchange between satellites and between satellites and the ground will become more frequent. Communication constellations need link scheduling, remote - sensing constellations need data transmission, and the future space - based computing network also needs in - orbit processing, inter - satellite forwarding, and task coordination.

This will increase the importance of space optical communication.

In the past, space optical communication was more regarded as a technical path to improve inter - satellite transmission capabilities. Now, it may become a key link for the space - based network to access the intelligent computing system.

If the data obtained by satellites in orbit cannot quickly enter the ground computing power network or be efficiently transferred in the inter - satellite network, it will be difficult to support scenarios such as real - time remote sensing, emergency response, low - altitude supervision, and ocean monitoring.

The faster the ground intelligent computing network is, the higher the requirements for the aerospace network will be. Otherwise, satellites may become a bottleneck for data to enter the AI system.

This will also extend industrial opportunities from satellite platforms and launch services to optoelectronic communication payloads, inter - satellite links, high - speed satellite - to - ground transmission, network scheduling software, and intelligent constellation operation and maintenance.

IV. The Low - Altitude Economy Is a Real - World Scenario of Integration of Communication, Sensing, Computing, and Intelligence

The low - altitude economy is another application direction worthy of attention in this document.

The document proposes that the deployment of inference computing power on the edge devices of 5G/5G - A networks, optical networks, IP networks, and new - type industrial networks should be explored to provide edge computing services integrating communication, sensing, computing, and intelligence for scenarios such as transportation, the low - altitude economy, manufacturing, and entertainment.

This shows that the development of the low - altitude economy cannot only focus on aircraft or airspace management. After scenarios such as drone inspection, low - altitude logistics, urban air traffic, and emergency rescue are truly implemented, a complete set of network capabilities is required, including communication, perception, positioning, computing, scheduling, and safety supervision.

Drone inspection requires stable video transmission and target recognition. Low - altitude logistics requires continuous communication and route scheduling. Urban air traffic requires real - time perception and conflict warning. Emergency rescue requires the rapid establishment of communication and computing capabilities in complex environments.

These requirements are difficult to be met by a single technology. 5G - A, Beidou, low - orbit communication, edge computing power, and the urban low - altitude supervision platform will play roles in different aspects.

The key for the low - altitude economy to move from pilot projects to normal operation is not only "being able to fly", but also being visible, manageable, adjustable, and having fast computing speed.

Therefore, the low - altitude economy will become a real - world scenario of integration of communication, sensing, computing, and intelligence. It requires both the low - latency and large - uplink capabilities of the ground network and the high - precision spatio - temporal reference provided by Beidou. In scenarios such as the open sea, mountainous areas, and emergencies, low - orbit satellite communication will also be a supplementary capability.

From the perspective of the industrial chain, low - altitude communication networks, Beidou enhancement services, edge computing nodes, low - altitude perception devices, supervision platforms, and flight data processing systems will all become basic capabilities that need to be improved before the large - scale development of the low - altitude economy.

AI will not change the aerospace industry out of thin air; it will only raise the requirements for infrastructure.

Next, the aerospace capabilities that can enter the network, computing power, and data systems will be truly needed in the AI era.

This article is from the WeChat official account "Star Movement Without Bounds", author: UniLym, published by 36Kr with authorization.