-- The Leap East Summit was held in Hong Kong from July 8 to 10. Dr. Xiang Yanping, Chairman of the Technical Committee, AI Scientist and Co-founder of AICT(Artificial Intelligent Interconnection Technology Co., Ltd.), delivered a speech at the summit’s main forum. Addressing the current bottlenecks and iteration trends of AI technology, Dr. Xiang comprehensively elaborated on the brand-new technical paradigm of high-precision physical AI based on underlying technical logic. He analyzed the technical path of digital-real integration through a four-layer full-stack architecture, pointing out the direction for technological iteration and track segmentation of the global AI industry.
Leap East Summit is co-hosted by the Ministry of Communications and Information Technology of Saudi Arabia and the Innovation, Technology and Industry Bureau of the Hong Kong. As the inaugural Asia edition of Saudi Arabia’s flagship tech event LEAP, the event aims to build bridges for technology, venture capital and enterprises between Asia and the Middle East, and serves as a core hub for linking hard-tech industries of the two regions.
High-Precision Physical AI Restructures the Technical Foundation of the Physical World
At present, general AI centered on large models focuses on virtual content generation based on text Tokens. However, its technical architecture is inherently disconnected from the physical world, plagued by drawbacks such as insufficient perception accuracy, excessive end-side latency, and incapability of closed-loop physical execution, making it hard to meet industrial demands of real-world scenarios.
Dr. Xiang holds that real-world scenarios including urban governance and smart transportation impose rigid constraints on safety, real-time performance and stability, while general AI with strong generalization capacity cannot resolve core pain points such as refined governance and full-domain security prevention and control. He stated: "General AI addresses demands for textual imagination, while high-precision physical AI tackles real-world demands. Without a technical foundation enabling high-precision perception of the physical world, AI empowerment for all industries will remain a castle in the air."
In his view, competition in the second half of the global AI industry is no longer merely a race over model parameters; instead, it hinges largely on technical adaptability tailored to the physical world. AICT put forward the concept of high-precision physical AI precisely to break technical barriers between virtuality and reality, rebuild a dedicated technical foundation adapted to the real physical world, and pioneer a core path for AI technology to break through application bottlenecks.
Four-Layer Full-Stack Architecture Creates a New Technical Paradigm for High-Precision Physical AI
Different from the design logic of general AI that separates algorithms and hardware, Dr. Xiang systematically unpacked AICT’s progressive four-layer architecture of high-precision physical AI: Token - AI Agent - AI Station - AIStation OS. He clarified that this system acts as the core carrier for seamless integration of high-precision physical AI algorithms, hardware, computing power and scenarios, establishing an innovative technical paradigm for physical AI.
Token serves as the cornerstone of this technical paradigm. High-precision physical AI conducts multi-modal structured encoding of real-scene information via Tokens, standardizing unordered physical data to address the fragmentation of real-world data and deliver accurate underlying data support for full-link intelligent computing.
AI Agent functions as the core decision-making hub. Trained on precise Token data and embedded with physical space constraint rules, it abandons the generalized reasoning mode of universal agents. It fits standardized deployment of unmanned devices and robots in real scenarios such as cities, transportation hubs and airports, realizing unattended precise scheduling, risk early warning and closed-loop disposal, and adapting to all scenarios requiring high-precision perception, computation and control.
AI Station represents a new integrated hardware-software carrier for AI. It integrates self-developed high-precision perception hardware systems such as Green Vision, large models, edge computing systems and distributed green energy storage-computing systems. By deploying perception, computation and decision-making locally, it achieves deep integration of hardware, algorithms and real physical scenarios. Each Green Vision node acts not only as a perception gateway, but also a spatial carrier for green power supply, distributed storage-computing and platform services. It continuously completes spatial status sensing, incident identification, local computing and intelligent collaboration, forming a distributed urban intelligent computing network.
AIStation OS is the top-layer architecture for full-domain collaboration. Through widely deployed AI Stations, it eliminates data silos and builds urban digital twins based on unified physical space. It enables coordinated scheduling across multiple industries including transportation, security and low-altitude economy, realizing technological upgrading from single-point intelligence to full-domain intelligence.
Divergence of Technical Paradigms Defines Future Tracks of the Physical AI Industry
In view of the iteration trend of global AI technology, Dr. Xiang put forward the view of track segmentation: "The global AI industry will split into two tracks in the future. Large models dominate virtual content, while high-precision physical AI leads the physical industry. This is the core leverage for Chinese tech innovators to achieve differentiated global competitiveness."
He pointed out that the core strengths of the new high-precision physical AI paradigm lie in closed-loop technology and native scenario adaptability. In contrast to overseas enterprises’ technical route that prioritizes algorithms while neglecting engineering and real-scene adaptation, Chinese high-precision physical AI has formed a complete technical chain covering underlying encoding, intelligent decision-making, hardware deployment and full-domain networking, which inherently fits diverse intelligent transformation demands of global physical industries.
"From a long-term industrial perspective, the ultimate iteration direction of AI technology lies in full-domain physical intelligence. Supported by the high scalability of the four-layer architecture, high-precision physical AI can flexibly adapt to scenarios of various scales with excellent technical reusability. As global industrial digital transformation accelerates, the new technical paradigm represented by high-precision physical AI will lead the development of the physical AI track and enable original Chinese technical concepts to continuously shape the evolution of the global AI industry," Dr. Xiang commented.
Contact Info:
Name: Jinna Wu
Email: Send Email
Organization: Artificial Intelligent Interconnection Technology Co., Ltd.
Website: https://www.aict.net/
Release ID: 89197806
In case of identifying any problems, concerns, or inaccuracies in the content shared in this press release, or if a press release needs to be taken down, we urge you to notify us immediately by contacting error@releasecontact.com (it is important to note that this email is the authorized channel for such matters, sending multiple emails to multiple addresses does not necessarily help expedite your request). Our dedicated team will be readily accessible to address your concerns and take swift action within 8 hours to rectify any issues identified or assist with the removal process. We are committed to delivering high-quality content and ensuring accuracy for our valued readers.
