Two rounds of financing secured within four months, Zhang Weijie intends to bring computing power "into the sky".

September 8, per exclusive disclosure from ChinaVenture, Deep Space Matrix, a space AI computing constellation company, has completed a 100-million-yuan angel round of financing, jointly led by CDH VGC and Lingge Venture Capital, with multiple front-line investment institutions participating as follow-on investors.

This is the second tranche of external funding Deep Space Matrix has obtained within several months of its establishment. In April 2026, two months after the company was officially founded, it completed a tens-of-millions-yuan seed round of financing. Meanwhile, its new round of financing is advancing smoothly.

In the process of building the "Chinese solution" for space AI computing power, Deep Space Matrix has finalized its transition from a broad entrepreneurial direction to a specific engineering plan. In July 2026, it launched the "Star Ring Plan", with a first-phase plan of about 210 satellites, aiming to build a low-orbit intelligent satellite network that integrates communication, remote sensing and computing.

The focus of the investment from this 100-million-yuan financing will still be R&D, engineering verification and system construction. According to Zhang Weijie, founder of Deep Space Matrix, the funds from this round will be mainly used for the R&D of core space AI computing technology, the system engineering verification of the "Star Ring Plan", and the ecological construction of the space AI industrial chain.

Investors Turn to Entrepreneurship

Build the "Chinese Solution" for Space AI Computing Power

Before founding Deep Space Matrix, Zhang Weijie had already accumulated rich experience in entrepreneurship and investment.

During his university years, he ran a business with his classmates, and entered the securities industry after graduation. Around 2012, judging that new financial business opportunities were emerging, Zhang Weijie chose to resign from Guotai Junan to start his own business, and finally achieved financial freedom.

It was investment that brought Zhang Weijie into the commercial aerospace industry. After investing in i-Space in 2018, he began to systematically engage in this industry, and founded Tianwen Times Fund, whose investment scope gradually covers sectors including rockets, satellites, TT&C, chips, materials, power systems and core components.

Long-term engagement with companies such as i-Space, Arrow Yuan Technology, Langxun Core Cloud, Hangju Technology, Huachi Kinetic Energy, Hangstar Transmission, Spacexys, Tianlian Aerospace, Aistron, Jiahang Future and Haiwen New Materials has continuously expanded and deepened Zhang Weijie's understanding of commercial aerospace, and shifted his core focus from individual technologies to industrial integration.

In his view, commercial aerospace has entered the 2.0 stage, whose core is to build system capabilities for the space AI era. The difficulty lies in how to integrate the capabilities of rocket, satellite, chip and other companies into a system, and further turn it into services that customers are willing to pay for.

As the 15th Five-Year Plan Outline lists aerospace as an emerging pillar industry, the state sets up a full-time agency to coordinate the development of commercial aerospace, and the national three-year action roadmap for commercial aerospace is accelerated for implementation, Zhang Weijie realized that the timing for entrepreneurship is ripe.

Zhang Weijie judges that the industrial capabilities of commercial satellite manufacturing, rockets, chips, materials and communications continue to accumulate, the engineering direction of recyclable rockets continues to advance, and AI also adds new technical options for the automatic processing of massive spatial data.

Based on the industrial judgment formed over years of investment, Zhang Weijie founded Deep Space Matrix in Beijing in February 2026, taking "serving the major strategic needs of the country and building a Chinese solution for space AI computing power" as the company's mission.

At the initial stage, Deep Space Matrix once considered providing satellites for large constellations. For a newly established commercial aerospace company, facing relatively clear demands and integrating itself into the existing supply chain is a path with clearer business boundaries.

After thorough investigation and repeated deliberation, they decided not to follow the path of simply supplying satellites, but to build their own constellation and directly target overseas customers and commercial scenarios. "If we only act as a supplier to a small number of large customers for a long time, it will be difficult for the company to grasp the final demand and build a complete commercial closed loop," Zhang Weijie explained.

The route change also redefines what the company needs to do. If it wants to build and operate its own constellation and be responsible for the final service results, it also needs to convert customer tasks into constellation requirements, arrange different satellites to perform their respective duties, so as to integrate the capabilities of remote sensing, communication, computing and different industrial links into one single system.

What Deep Space Matrix hopes to undertake is exactly such a general system role. It does not require all components to be manufactured by itself, but to define the entire system first and be responsible for the long-term sound operation of the final product, positioning itself more as a Chinese industrial chain organizer for private space AI constellations.

This ambitious goal is very challenging, which puts higher requirements on the team. At present, the company has an academician of the International Academy of Astronautics and a national meritorious chief satellite designer as its chief scientist, and the team covers fields including general satellite design, space-based intelligence, manned spaceflight, space station, AI chips, and heterogeneous supercomputing.

It is precisely because of the recognition of Deep Space Matrix's comprehensive strength that investment institutions such as CDH VGC and Lingge Venture Capital chose to make decisive investments shortly after the company was founded.

"We are highly optimistic about Deep Space Matrix's potential to become a full industrial chain organizer for space AI," Wang Lin, founding partner of CDH and managing partner of CDH VGC, believes that space AI is not a single-point technical opportunity, and new industrial opportunities will arise in sectors ranging from transportation capacity, satellite platforms, chips, to communications, energy, thermal control, materials and manufacturing. But what is truly scarce is an enterprise that can define the system from the top level and organize the excellent capabilities of the industrial chain at the same time.

Lingge Venture Capital pays more attention to the implementation capability of Deep Space Matrix. In the view of Lingge Venture Capital, the final outcome of space AI is grand enough, but what really matters is to proceed from practical constraints, and split the long-term goal into a system construction path that is "capable of design, verification and iteration". "The Star Ring Plan is exactly such a solution that takes into account both long-term potential and phased implementation capabilities."

Start with a Low-Orbit Constellation

Launch 210 Satellites into Space

If the "Star Ring Plan" is only understood as building another low-orbit constellation, it is easy to ignore what Deep Space Matrix really wants to sell.

Zhang Weijie summarizes the revenue models of the satellite industry into three categories: "The first category is to sell satellites; the second category is to sell data; the third category is to sell services." Deep Space Matrix, which has decided to directly connect with end customers, hopes to continuously provide remote sensing and judgment results in the long run.

Before its own satellites form stable service capabilities, Deep Space Matrix plans to first integrate the existing multi-source remote sensing data in the market to provide services for early demands and verify the scenarios. As its own satellites are gradually put into use, it will further improve the revisit frequency, data acquisition capability and service stability.

"Commercial aerospace must first be commercial. We will not launch a large number of satellites into space first and then wonder who to sell them to. Instead, we hope to develop our business while providing services, so that our business and the constellation can grow together," Zhang Weijie added.

Although it has only been established for a few months, the company is already advancing its commercial implementation. At present, it is docking demands in key overseas markets under the Belt and Road Initiative such as ASEAN and Africa, with early focus on scenarios that require massive spatial data and high timeliness, including energy facility monitoring, agriculture and forestry monitoring, natural resource survey, and emergency response.

For these potential customers, a remote sensing image is often not the final product. Taking offshore oil and gas facility monitoring as an example, what customers really care about is where anomalies occur, whether leakage is possible, when the change takes place, etc., so as to take actions based on the obtained detection and judgment results.

This is also the reason why Deep Space Matrix's AI constellation "Star Ring Plan" is designed as a network integrating communication, remote sensing and computing. The first phase of about 210 satellites will include different nodes for remote sensing perception, communication relay and computing. The perception nodes acquire data, the communication nodes complete transmission, and the computing nodes are responsible for processing and analysis, finally forming a task chain from data acquisition to result delivery.

Once the product becomes such a system, Deep Space Matrix cannot rebuild all capabilities including rockets, chips, communications, materials, energy and manufacturing internally. The professional companies that Zhang Weijie invested in and contacted with when he laid out the commercial aerospace industry in the past have therefore become the realistic basis for him to choose "open integration".

Different from SpaceX's highly vertically integrated development model, Deep Space Matrix chooses to explore a path that is more in line with China's industrial foundation, and build a "Chinese solution" for space AI. In short, it will be responsible for the general system design and core capability construction, and at the same time, in key links such as transportation capacity, computing power, energy, communication, thermal control, materials and manufacturing, it will promote the "Star Ring Plan" jointly with industrial partners through joint R&D and open collaboration.

Using the existing launch missions of partners to carry out piggyback verification is one of the most specific actions on this path at present. According to Deep Space Matrix's plan, part of its core capabilities will be prioritized to be carried on partners' launch missions for product testing in a real on-orbit environment, before entering the first batch of satellite projects, and gradually promote the deployment of about 210 satellites in the first phase.

Behind this series of choices, there is also the consideration of cost control. Zhang Weijie believes that the traditional aerospace system has long emphasized high reliability of a single satellite. If the same engineering method is directly copied to the scale of hundreds or thousands of satellites, the cost of a single satellite may not be able to support commercial constellations. "The company therefore hopes to reduce costs through general integration and reducing redundant systems."

For Deep Space Matrix, the issues that need to be addressed in the future are whether the capabilities scattered in different links of commercial aerospace can be organized into a service that can be delivered stably, and whether customers are willing to pay for this service for a long time. The former determines whether this system can operate, while the latter determines whether it can become a viable business.

In the long run, starting from today's spatial intelligent services, Deep Space Matrix hopes to gradually move towards a larger-scale space AI computing power service base. "Remote sensing satellites are the 'eyes', the communication network is the 'nerve', and increasingly powerful space computing nodes in the future are the 'brain'. Users only need to put forward tasks, and the system will automatically call spatial data, communication networks and computing resources to complete them."

This article is from the WeChat official account "ChinaVenture", author: Lu Zhigao, published with authorization from 36Kr.