The HICOOL 2026 Global Entrepreneurship Competition concluded its finals in Beijing, with Feidu Tech's spatial intelligence "Zhengrong Model" securing the first-place award. According to public records, this year's event drew entrepreneurial ventures spanning artificial intelligence, biomedicine, new energy, and high-end equipment manufacturing. The accolade underscores the jury's recognition of Feidu Tech's advancements in spatial intelligence technology development and its practical industry applications. Now in its seventh consecutive year, the HICOOL competition has attracted over 44,000 projects from 168 countries and regions.
Feidu Tech's entry, the Zhengrong Model, is engineered for industrial-grade spatial intelligence scenarios. Diverging from AI products focused on content generation or consumer-facing experiences, this model emphasizes the generation, comprehension, reasoning, and decision-making capabilities for three-dimensional spaces and physical environments. It is designed to address complex use cases such as urban governance, flood prevention, and embodied intelligence. The model has completed filing with the Cyberspace Administration of China, according to the company.
Its technical architecture covers 3D spatial data generation, semantic understanding of spatial environments, simulation and reasoning of physical laws, and scenario-based decision output. This is structured as a five-layer technology stack encompassing appearance, geometry, semantics, physics, and simulation. By integrating physical principles with spatial data, the company aims to push AI beyond mere spatial recognition, enabling it to comprehend, project, and support decision-making processes.
The model is already deployed across multiple sectors. In urban construction, Feidu Tech is involved in a city-level LOD1.3 3D modeling project. This initiative employs a technical route where 2D image analysis guides 3D point cloud reconstruction, fusing multi-source spatial data for automated modeling. The company states this approach boosts production efficiency by over 80% compared to traditional methods, with outputs applicable to urban flooding simulations, city renewal projects, and 5G signal propagation modeling.
Within the hydraulic sector, the Zhengrong Model supports a watershed flood and flash flood forecasting initiative. By integrating meteorological, rainfall, and hydrological data, it facilitates flood routing simulations, risk evaluations, and the optimization of response plans. In embodied intelligence, Feidu Tech has developed a spatial intelligence platform for robots in public venues, offering foundational capabilities like spatial mapping, environmental interpretation, and task planning for robotic operations.
On the international front, Feidu Tech's urban development intelligent agent has been integrated into Abu Dhabi's city-wide spatial intelligence platform. This deployment supports long-term urban planning simulations, assessing the impact of construction proposals on real estate, transportation, and environmental factors. The project adopts an integrated model combining infrastructure, engine, data, and operations, serving as a reference case for the company's expansion into the Middle East and other overseas markets.
Founded in 2016, Feidu Tech is an early player in China's spatial intelligence infrastructure space. The company established its AI research institute in 2021 and released a spatial large model in 2024, transitioning its focus from 3D space digitalization to spatial understanding and decision-making. Today, its operations span over 50 countries and regions, with strategic emphasis on the Middle East, Southeast Asia, Europe, and North Africa. The company has completed a Series B+ funding round, accumulating hundreds of millions of yuan in total financing.
Looking ahead, Feidu Tech plans to advance its research into spatial intelligence and world models, refine its product offerings for urban, hydraulic, and robotics sectors, and accelerate both overseas expansion and ecosystem development. As AI evolves from language and image comprehension toward engagement with the physical world, the ability to standardize spatial intelligence products, replicate them across industries, and consistently reduce project delivery costs will be pivotal in validating the commercial viability of companies in this space.