According to the Zhitong Finance APP, the 2025 World Artificial Intelligence Conference (WAIC) will be grand in scale, with numerous participants focusing on the latest technological advancements and commercial applications in the field of artificial intelligence. This article provides an overview of the conference, covering the robotics industry, large models and AI applications, hardware innovations, as well as market and industrial ecology. Recently, there has been a significant rebound in the AI sector, with most companies benefiting, achieving up to 50%+ returns, while the worst performers saw a decline of 10%-20%; influenced by the unexpected events at WAIC (difficult ticket purchases, registration closing after July) and GPT-5, the market shows sustainability, with institutions adopting a high-low cutting strategy, maintaining a simulated position of 90% high holdings. The A-share computer/AI application sector is in a beta market, with rapid rotation of hotspots and a lack of sustained alpha returns; corporate performance growth is stable (most revenue growth is 10%+), and AI revenue scale remains small, not yet reaching an explosive stage.
<h2>I. Basic Overview of the Conference</h2>
The 2025 World Artificial Intelligence Conference (WAIC) will be grand in scale, with numerous participants focusing on the latest technological advancements and commercial applications in the field of artificial intelligence. Among them, the robotics exhibition area stands out, with the number of exhibitors, exhibition area, and audience numbers all reaching historical highs — the area of the robotics exhibition hall increased by about 40% compared to last year, with over 40 exhibitors of biped humanoid robots, and a total of about 60 industrial robot manufacturers, doubling from last year; the number of dexterous hand manufacturers reached 9, and the number of rotary joint module manufacturers significantly increased, with integration becoming an industry trend.
<h2>II. Exhibition Area Features and Technical Highlights</h2>
Significant technical differences in sub-areas
H1 Exhibition Area: Dominated by major companies such as Baidu and Alibaba, the booth area has expanded, with diverse application displays. The progress of edge-side models has been rapid, such as the optimization of intelligent interaction for automotive edge-side models and the improvement of document processing efficiency for PC-side linear edge models, marking a breakthrough in AI from 0 to 1 in practical scenarios.
H4 Exhibition Area: Focused on emerging technologies and cutting-edge applications, showcasing innovations in niche areas such as AI in Model-Based Systems Engineering (MDSE) design, shortening the design cycle of complex industrial systems and laying the groundwork for industrial transformation.
Robotics Exhibition Area: Leading manufacturers (Zhiyuan, Yushu, Fourier, Leju, etc.) have concentrated resources, with large booth areas, having entered the stage of commercial mass production. The number of core component manufacturers such as dexterous hands and rotary joint modules has surged, with significant improvements in technological maturity.
<h2>III. Robotics Industry: From Technological Breakthroughs to Commercialization</h2>
1. Product and Scenario Progress
Industrial Scenarios: The wheeled robot developed by Zhiyuan in collaboration with DeMa Technology achieves autonomous sorting of packages with an error of ≤5 cm; Feilong Technology's robot, equipped with self-developed torque sensors, can precisely grasp fragile items such as glass and chips; Kepler's bipedal and wheeled robots have strong endurance, focusing on material handling; the Galaxy General wheeled chassis has high mobility efficiency, improving factory material handling efficiency by 3 times compared to manual labor.
Commercial Scenarios: The Galaxy General supermarket sorting robot performs stably, running for 8 consecutive hours with zero errors.
Rehabilitation Medical Scenarios: Fourier Medical's exoskeleton robot has entered dozens of hospitals in China, assisting in the rehabilitation of individuals with walking disabilities; its biped humanoid robot is used for cognitive impairment and autism rehabilitation companionship, with the new generation of family emotional companionship robot GR 3 (shaped like &#34;Baymax&#34;) making its first public appearance
Industry Data: The expected shipment volume of leading manufacturers in 2025 is about 5,000 units at most, with most between 500-2,000 units, totaling over 10,000 units.
2. Challenges and Technical Directions
Currently, most robots still require manual control or lack sufficient autonomy, needing to enhance the &#34;brain&#34; intelligence level (multi-dimensional perception, autonomous decision-making, complex task decomposition). Domestic companies still lag behind overseas companies like Tesla and bigger in terms of autonomy.
Technical Trends: Integration of rotary joint modules, improvement of dexterous hand precision; in the core component links, encoders (eddy current encoders, rotary transformers), reducers (planetary RV cycloidal pinwheel actuators), lead screws, motors, etc., have become key breakthrough directions, involving companies such as Yapu Co., Ltd., Shuanghuan Transmission, Hengli Hydraulic, and Boke Co., Ltd.
<h2>4. Large Models and AI Applications: From &#34;Quantity Competition&#34; to &#34;Effective Implementation&#34;</h2>
1. Development Trends of Large Models
The &#34;Hundred Model War&#34; of 2023 has shifted towards a converging pattern, with core manufacturers continuously iterating: SenseTime released the &#34;Daily New 6.5&#34; multi-modal reasoning large model, employing a design of interleaved text and image thinking chains, enhancing reasoning capabilities while reducing costs by 70%; mini max launched the open-source large-scale hybrid architecture reasoning model M1, achieving breakthroughs in video large models and intelligent agent fields.
Domestic startups have achieved breakthroughs in multi-modal reasoning under limited computing power, especially in long video understanding, enterprise services, and mobile applications; the overseas version K2 of the Kimi large model has enhanced reasoning capabilities through architectural optimization, becoming an alternative option to mature overseas models.
2. Application Scenarios and Industry Focus
Industry-wide service agents: The demand for computing power (especially reasoning computing power) has surged, with large models focusing on H parameter capabilities (coding, tool invocation, etc.), penetrating from general applications into industries.
Vertical Breakthroughs: In the medical field, IFLYTEK is deeply engaged in AI diagnosis, while the learning machine market maintains high growth relying on &#34;AI personalized tutoring&#34;; in the financial sector, Jinyue Xingchen collaborates with brokerages on the &#34;Question Finance Agent,&#34; achieving automation in financial report analysis and research report breakdown; in the industrial sector, Baoxin showcases AI applications for energy efficiency optimization in the steel industry.
Benchmark Cases in Industrial Internet: Deep Yuan Artificial Intelligence and Haier Group's Kaos COSMOPlat joint booth attracted attention, showcasing the world's first L4 level intelligent body master system MasterAgent, with both parties deeply cooperating around core scenarios of the industrial internet, achieving intelligent recruitment, full-process automation of customer service, and intelligent scheduling of enterprise workflows, further unleashing the potential of AI entities in this field.
<h2>5. Hardware Innovation and Corporate Dynamics</h2>
1. Highlights of Intelligent Hardware
Out of the Door Ask Tik Note: As a personal information terminal, it supports recording transcription summaries, mind mapping, and voice interaction, building a personal knowledge base.
Kuaishou Lingdong Canvas: A real-time collaborative AR creation platform, with a video generation cost of ≤10 yuan for 10 seconds, rapidly transforming from the laboratory stage into a productivity tool
Titan Dynamics Technology novels NAVOS: Overseas marketing agent, relying on large models to provide full-link services, solving high localization costs and talent scarcity issues.
IFLYTEK multilingual AI smart screen: Transparent screen supports cross-language barrier-free communication, paired with the newly released AI recording pen to enhance scene coverage.
2. Enterprise Product Iteration
Kingsoft released WPHAI 3.0, transitioning the Lingxi product from WPS toolbar interaction to an independent entry, moving towards autonomous intelligent agents.
SenseTime's embodied intelligent platform: Expanding based on the autonomous driving world model, utilizing 100,000 3D assets to enhance robot vision and motion control capabilities.
<h2>VI. Market and Industrial Ecology</h2>
1. Market Trends and Investment Trends
Recently, there has been a significant rebound in the AI sector, with most companies benefiting, reaching up to 50%+, and the worst at 10%-20%; influenced by the unexpected WAIC event (ticket prices hard to buy, reporting stopped after July) and GPT-5, the market shows sustainability, with institutions adopting high-low cutting strategies, maintaining a 90% position at high levels.
The A-share computer/AI application is in a beta market, with rapid rotation of hotspots and a lack of sustained alpha returns; corporate performance growth is stable (most revenue growth is 10%+), and AI revenue scale is still small, not yet reaching an explosive stage.
2. Industrial Ecology and Policy Support
Changes in the structure of participating companies: Data labeling vendors (such as Aopeng China) have shifted their clients from model vendors to enterprises/governments; application vendors (Youdao, Shenxinfu, etc.) are growing in size, while pure model vendors are decreasing.
Local government layout: The Yangtze River Delta region centered around Shanghai has become the core area of the AI industry, attracting companies to settle through policies such as computing power vouchers and application subsidies, utilizing local computing power resources to build industrial clusters.
China-US ecology: The entrepreneurial enthusiasm remains high, with active industrial metabolism, and computing power is a key investment area, with outstanding talents focusing on specific categories such as AI video, coding, and for science.
<h2>VII. Summary and Outlook</h2>
WAIC 2025 showcased the accelerated penetration of AI from technology research and development to commercial landing: The robotics industry is moving from &#34;showing off technology&#34; to practicality, large models are shifting from &#34;parameter competition&#34; to &#34;value creation,&#34; and hardware and applications continue to reduce costs and get closer to people's livelihoods. Despite challenges such as agent homogeneity and insufficient robot autonomy, government support, active capital, and technological iteration are jointly driving the industry forward. AI, as a young and vibrant industry, has broad prospects for the future.
Note: This article is based on the reflections of analysts from various brokerages such as CICC and Guotai Junan after attending WAIC. For complete reflections or image support, please contact relevant institutions or refer to the official channels of the conference

002230

00020

The 2025 World Artificial Intelligence Conference (WAIC) is grand in scale, with the robotics exhibition area performing outstandingly, setting historical highs in the number of exhibitors, exhibition area, and audience attendance. The area of the robotics exhibition hall increased by about 40% compared to last year, and the number of bipedal humanoid robot and industrial robot manufacturers has doubled. There is a noticeable rebound in the AI sector, with most companies experiencing revenue growth, indicating sustainability in the market. The conference focuses on the latest technologies and commercial applications of artificial intelligence, with significant technological differences in the exhibition area, showcasing diverse applications and innovations

- The 2025 World Artificial Intelligence Conference (WAIC) showcased significant advancements in AI technology and applications, particularly in robotics, large models, and hardware innovation.  
- The AI sector is experiencing a rebound, with many companies reporting earnings growth of over 50%, driven by events like WAIC and the anticipated impact of GPT-5.  
- Despite the rapid developments, the AI revenue scale remains small, and the industry faces challenges such as a lack of sustained alpha returns and the need for improved robot autonomy.

2025 WAIC Comprehensive Observation: From L4 Level Intelligent Agents to Robots that Can Deliver Packages, AI Has Truly Landed!