09,29,2026 17 views
Recently, Guanghetong has completed a strategic investment in Beijing Feidu Technology Co., Ltd. (hereinafter referred to as "Feidu Technology"), further expanding into the fields of spatial intelligence and world models, and expanding the AI+robot industry ecosystem. Feidu Technology has long focused on the field of space intelligence, committed to developing a full chain and full stack space intelligence infrastructure for "generation, understanding, reasoning, and decision-making" in the physical world, deeply empowering thousands of industries.
Key variable
An increasingly prominent bottleneck for AI's transition from the digital world to the physical world is the acquisition of high-quality real-world data. Physical AI applications such as robots and autonomous driving require a large amount of data containing spatial relationships, physical laws, and behavioral trajectories for training. However, data collection in real-world environments is often costly, time-consuming, and difficult to fully cover dangerous scenarios, extreme working conditions, and low-frequency events.
The world model and spatial intelligence provide new technological paths for this problem. By generating and reconstructing the real physical environment, and combining spatial relationships, physical properties, and dynamic data, a computable, interactive, and replicable digital world can be constructed, allowing some data generation, training, testing, and predictive inference to move from the real environment to the virtual space.
In 2026, the World Economic Forum listed World Models as one of the "Top 10 Emerging Technologies", stating that they can learn the operating rules of physical environments from multi-source data such as videos, sensors, and text, helping AI understand and predict changes in the real world.
This technological path has extended to multiple real-life scenarios.
Robots can generate data, train repeatedly, and validate algorithms in virtual environments;
Autonomous driving can conduct simulation testing based on high-precision road and traffic environments;
Games can utilize real spatial data and 3D scene capabilities to construct large-scale virtual environments;
At low altitudes, the external environment such as urban buildings, terrain, airspace, and flight routes can be digitized first, and then dynamic data such as aircraft, weather, and operating rules can be overlaid for flight scheduling and risk deduction;
In the field of urban governance, it is also possible to simulate processes such as water flow, rainfall, and flood routing, providing a basis for prediction, warning, and decision-making assistance.
From 'collecting the real world' to 'training, testing, and deducing in the digital world', space intelligence is changing the way data is obtained and validated for physical AI and complex industry applications. This is also an important reason why Guanghetong continues to pay attention to the long-term value of world models and space intelligence.
Core Accumulation
In this direction, Feidu Technology has accumulated long-term experience in data resources, technological capabilities, and industrialization practices, which is also an important consideration for Guanghetong's investment.
Feidu Technology independently builds 3DT spatial data standards and data engines, supporting processing of over 100 multi-source heterogeneous spatial data formats, and accumulating billions of real spatial data corpora to build a data foundation for spatial intelligence model training and application. At the same time, the company continues to promote the construction of domestic 3D computing capabilities, develops its own DTS rendering engine, and supports real-time rendering of large-scale data in the information and innovation environment.
On this basis, Feidu has launched the "Zhengrong" spatial intelligence model, which further extends its capabilities from spatial data processing to spatial generation, understanding, simulation, and decision-making. At present, the company has served about 2000 enterprise clients, formed over 100 deep industry solutions, and expanded its business to more than 20 countries and regions around the world, with practical applications in urban governance, low altitude economy, robotics, and other scenarios.
For embodied intelligence, Feidu is still building a training environment that covers SimReady asset generation, physical simulation, and data synthesis capabilities, providing support for robot training and algorithm validation.
Song Bin, co-founder/director of Feidu Technology, said:
Space intelligence is accelerating from technological exploration to industrial applications. Feidu has been continuously investing in real space data and space intelligence technology, promoting AI to further understand, simulate, and deduce the physical world. The acquisition of strategic investment from Guanghetong this time is a recognition by industry partners of the long-term development value of space intelligence. In the future, Feidu will continue to deepen technological innovation and promote the landing of space intelligence in more real-life scenarios. ”
Continuous Layout
Chen Shijiang, Senior Vice President and Secretary of the Board of Guanghetong, stated:
AI is moving further from perception and generation to understanding, simulating, and making decisions about the real world. Feidu Technology has accumulated long-term experience in spatial data, spatial intelligent models, and industrial applications, which is highly consistent with our judgment on the development direction of Physical AI. This investment is another forward-looking layout of Guanghetong around the AI+robot strategy. In the future, we will continue to focus on key technologies and innovative forces in the evolution of the industry, and further expand the AI+robot industry ecosystem. ”
Guanghetong continues to pay attention to different key technological links around the evolution of AI+robot industry. Previously, the company had strategically invested in areas such as end-to-end AI computing power; This time, we will further extend our investment vision to spatial intelligence and world models, and continue to improve our investment layout in AI+robot cutting-edge technologies.
