51WORLD(06651)Formally reached a strategic cooperation in the aerospace field with Envion Intelligence.

51WORLD (06651) announced that the company has recently reached a strategic cooperation agreement with SpaceTime Intelligence Technology Co., Ltd. (SpaceTime Intelligence) in the aerospace field. SpaceTime Intelligence is a leading domestic commercial aerospace enterprise focusing on the overall research and development and in-orbit operation of Earth observation remote sensing satellites. It is also the largest commercial aerospace entity in the southwest region with the capability to independently build and operate a remote sensing satellite constellation. The two companies will cooperate in the collection of high-precision satellite remote sensing data and the reconstruction of "physically usable" data assets, and jointly customize the world's first commercial remote sensing satellite designed specifically for physical AI training and simulation applications - Earth Clone Star "ECS-1." This cooperation marks the first deep collaboration between the Chinese physical AI industry and commercial aerospace in satellite payload construction. Earth Clone Star "ECS-1" is the world's first commercial remote sensing satellite customized for the specific requirements of physical AI data, with several leading industry indicators. The ECS-1 satellite is equipped with an optical imaging payload with a ground sampling distance (GSD) of 0.5 meters, and an imaging width on the ground of 60 km; the satellite uses an advanced stereoscopic push-broom imaging system with the ability for stereo mapping, enabling the generation of high-precision digital surface models (DSM) and digital elevation models (DEM) in a single pass. In comparison, traditional commercial optical remote sensing satellites typically require multiple track crossings and joint calculation to achieve similar accuracy in three-dimensional terrain reconstruction. In terms of comprehensive efficiency: (1) Significant optimization in three-dimensional terrain reconstruction efficiency: Leveraging a single-track stereoscopic imaging configuration, the satellite can complete regional stereo mapping in a single pass, significantly reducing the operation cycle for regional three-dimensional reconstruction compared to the traditional single-line array optical satellite multi-track stitching modeling mode; (2) Data processing efficiency multiplied: Combining the company's 51WORLD MODEL world model with the 4DGS (four-dimensional Gaussian splatter) AI processing pipeline, the end-to-end cycle from satellite capture to deliverable scene is reduced from the industry's traditional days to hours; (3) Substantial reduction in data acquisition cost: The comprehensive acquisition cost of real-world three-dimensional data at the city to provincial scale can be reduced by approximately 90% compared to traditional aerial oblique photography. In terms of project progress, ECS-1 has recently completed the construction plan review and officially started the bidding process, with plans to complete satellite launch networking by the end of 2027. For a long time, the commercial value of commercial satellite data has been difficult to fully realize, mainly due to certain technical challenges in connecting original remote sensing images with the training and simulation needs of physical AI. As a leading company in the field of physical AI simulation in China, the company has the full-stack capability to rapidly transform original geographic remote sensing data into "physically usable" simulated scenes. In this cooperation, the data collection at the satellite end and the physical AI scene reconstruction will form a direct pipeline, evolving from "perceiving the Earth" to "understanding the Earth." Based on the company's self-developed physical AI engine and 51WORLD MODEL world model, the company undertakes the core functions of "understanding" and "generating": through multimodal perception and large-scale spatial semantic learning, the model can encode ECS-1 and existing satellites, aviation, and ground-based heterogeneous data into a unified format, achieving cross-scale reconstruction from pixel-level images to semantic digital worlds. Utilizing the 4DGS (four-dimensional Gaussian splatter) neural rendering pipeline, the original stereo images are sequentially matched and DSM/DEM generated, sparse point cloud rebuilt and densified, 4DGS neural rendering, full-scene semantic segmentation and element-level annotation, standardized semantic system mapping, and binding of physical properties (such as friction, reflectivity, passability, carrying capacity, etc.), are processed in an automated pipeline, resulting in "understandable, interactive, and trainable" ready-to-use "physically usable" data assets, directly available for downstream physical AI applications. In addition, the two parties have planned for the SAR satellite ECS-2, equipped with a full-polarization synthetic aperture radar and InSAR deformation monitoring capabilities, with imaging width of 40km, achieving all-weather observations throughout the day by penetrating through clouds and fog; when cloud cover causes the optical imaging of ECS-1 to fail, ECS-2 collaboratively supplements the data collection in critical areas to ensure uninterrupted data acquisition. The data from the ECS-1 satellite can provide stable underlying data support for the company's full range of physical AI businesses, expanding commercial aerospace application scenarios, empowering digital twins, autonomous driving simulations, Siasun Robot&Automation training, low-altitude economy, geological emergencies, natural resources and major engineering management, among other fields, bridging existing capability gaps and gaining cost advantages. In the future, it can support the company's deployment of space computing power, extending the ground computing system to near-earth orbit, and establishing an integrated intelligent perception and application closed loop between sky and earth.