-+ 0.00%
-+ 0.00%
-+ 0.00%
Cathay Pacific Haitong: AI's move to physical world data services ushered in new opportunities
Share
Listen to the news

The Zhitong Finance App learned that Cathay Pacific Haitong released a research report saying that physical AI is developing at an accelerated pace, demand for high-quality data continues to grow, and data supply has become a key factor limiting model development. As the data service market expands, enterprises with data resources, large-scale production and platform service capabilities are expected to benefit. Focus on the two major directions of real data acquisition and processing, simulation synthesis, and spatial data. 1) Real data acquisition and processing: Companies with real scene resources, multi-modal data accumulation and mature acquisition and processing capabilities are expected to benefit; 2) Simulation synthesis and spatial data: enterprises with the advantages of high-fidelity physical engines, 3D scene generation, and spatial data resources are expected to benefit.

Cathay Pacific Haitong's main views are as follows:

The development of physical AI and world models is accelerating, and data demand continues to grow

AI is expanding from the digital world to the physical world. Model capabilities are gradually expanding from environmental perception to physical state prediction, robot motion generation, and complex task execution. Manufacturers such as Google, NVIDIA, and Tencent are speeding up the layout of world models and promoting the accelerated integration of VLA and world models. Model training data is also extended from traditional text, images, and videos to multi-modal interactive data such as actual operation trajectories, joint states, force, and tactile feedback. In terms of data supply, the Internet has advantages of open data, unstructured data, simulated synthetic data, and real machine data to jointly support model training. The global physical AI and world model data service market is about 17.75 billion yuan in 2025, and is expected to reach 131.23 billion yuan in 2030. The CAGR is about 49.2% in 2025 to 2030. The Chinese market is expected to grow from 8.28 billion yuan to 64.99 billion yuan during the same period, with a CAGR of about 51.0%.

Domestic and foreign manufacturers are speeding up the layout, and specialized data service systems are gradually being formed

Overseas manufacturers mainly rely on original AI data services, expert networks, and model evaluation capabilities to expand into the field of physical AI. Among them, Scale AI launched the Physical AIDATA Engine, which lays out robot remote operation, first-person data collection, and multi-modal labeling; Surge AI, Mercor, Micro1, and Handshake AI expand services around expert training data, model evaluation, enhanced learning environments, and real scene data. The domestic optical wheel intelligence focuses on physical AI data and simulation evaluation infrastructure to form a full-stack product system covering human behavior data, simulation synthesis, simulation evaluation, and deployment feedback.

Data resources, cost efficiency and platform ecology build long-term competitive barriers

In terms of data resources, mature physical models require tens of millions of hours of real data, and currently the world's high-quality real data is only hundreds of thousands of hours, and enterprises with stable data sources, rich scene coverage, and standardized quality control capabilities are more advantageous; in terms of cost efficiency, real machine acquisition costs are high, and methods such as UMI, EGO and simulation synthesis are expected to reduce data production costs, and AI agent assisted collection is also expected to reduce manual input and improve data output efficiency; in terms of platform ecology, data services are moving from a single collection and labeling to the entire chain of data management, model training, evaluation and deployment feedback Road service upgrades, enterprises with cross-entity adaptation, platform integration and industrial chain collaboration capabilities are expected to enhance customer stickiness and continuous service capabilities.

Risk warning: The development of the physical AI industry falls short of expectations; progress in world models and simulation technology falls short of expectations; changes in data collection technology routes; intensification of industry competition and commercialization falling short of expectations; data security, copyright and compliance risks.

Disclaimer:Webull uses external vendor Google Translation Service for news translations where we endeavour to ensure these are correct, however, we recommend that you please double-check this information accordingly. Webull is not responsible for translation errors or issues.
What's Trending