<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/a02a9853-a15d-46ca-a79d-67213024219e.jpeg?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/a02a9853-a15d-46ca-a79d-67213024219e.jpeg"/>
Author | Chai Xuchen
Editor | Wang Xiaojun
Ant Group, which has not yet personally entered the field of robotics, has already made early bets on &#34;potential stock&#34; players in the industry.
On December 6, at Lingnan Stadium of The Chinese University of Hong Kong, there were no flawless robot performances under the spotlight; instead, a group of young people waited by the grass, stone steps, and suspension bridge.
A quadruped robot was attempting to walk across a swaying suspension bridge, its gait somewhat hesitant; on the other side, a robotic arm repeatedly tried to grasp a plastic bottle on the grass but missed several times. This scene seemed clumsy, but in the eyes of Liu Yunhui, chairman of the event expert committee and academician of the Hong Kong Academy of Engineering, it was more exciting than any carefully choreographed demonstration—because there were no remote controls, no pre-set scripts, and all actions relied on the robot's own perception, decision-making, and execution in a real, chaotic outdoor environment.
This was the final of the fifth ATEC Technology Elite Competition. As a key force behind the event, Ant Group aimed to bring robots out of the greenhouse of the laboratory and into the harsh real world.
It is understood that the competition for the first time placed the entire competition field in an outdoor natural terrain, with scoring rules clearly encouraging operations without remote control, promoting the evolution of robots from relying on remote control as &#34;tools&#34; to autonomous decision-making agents.
Achieving &#34;no remote control&#34; means that robots must independently complete the entire closed-loop process from perception, analysis to decision-making and execution in a real environment full of uncertainties. Any mistake in any link could lead to task interruption, which poses extremely high requirements for the robot's perception robustness, decision-making intelligence, and system stability.
&#34;This competition aims to answer a core question: Can robots truly step out of the laboratory and adapt to our complex world?&#34; Liu Yunhui stated, &#34;We hope to promote the transition of robots from 'demonstrably feasible' to 'reliably applicable' through extreme challenges.&#34;
Therefore, the competition scene was deliberately set in an unstructured outdoor environment: variable lighting, irregular terrain, and leaves blown by the wind. Robots needed to complete four major tasks: &#34;orienteering,&#34; &#34;suspension bridge crossing,&#34; &#34;autonomous watering,&#34; and &#34;garbage sorting.&#34; Behind this are countless complex problems and precise calculations.
&#34;Perfect algorithms debugged in the laboratory will encounter countless surprises in the real environment,&#34; admitted Zhu Chengrui, captain of the Wongtsai team from Zhejiang University. In the &#34;autonomous watering&#34; segment, the robot not only had to identify the watering can and flowers but also had to plan its path and control its strength while moving; any error in any link could lead to task failure.
Zhu Chengrui said that it is still too complex to truly integrate robots into human daily life. &#34;For robots, even a small pillow placed in the wrong direction requires them to re-understand.&#34; Academician Liu Yunhui pointed out that this tests the three core capabilities of robots: walking, operating, and modifying the environment, and in real scenarios, the difficulty increases exponentially from &#34;walking&#34; to &#34;taking action.&#34;
Putting the competition venue in mountainous areas, grasslands, stone steps, and suspension bridges allows robots to face real-world disturbances. Every challenge designed is meant to expose their true weaknesses in collisions. This is exactly what Ant Group, as the initiating unit, wants to see.
The head of the Technology Strategy Department at Ant Group bluntly stated, &#34;If the problems are not real, they will not lead to genuine technological advancements. Only 'real problems' can inform the industry about what needs to be broken through next.&#34;
Behind this event, dubbed the &#34;Real World Extreme Challenge,&#34; lies Ant Group's greater ambition in the field of embodied intelligence—through a competition that touches on technological pain points, they aim to discover and bet on more top-tier robotic players like Yushu and Zhiyuan.
Ant Group's keen interest in this event stems from the fact that embodied intelligence is on the brink of explosion but is currently hindered by three major bottlenecks: environmental perception and cognition, intelligent decision-making and response, and the capacity of hardware and computing power. This compels young participants to seek breakthrough solutions, and it is precisely this extreme testing in the real world that provides them with a clearer direction.
Looking back, if Yushu Technology's founder Wang Xingxing opened the commercial era of quadruped robots with XDog, then what Ant Group is currently seeking is the potential force that can elevate embodied intelligence from &#34;Demo&#34; to &#34;Day One&#34; (usable from day one).
Currently, in the digital world, Ant Group's Lingguang and AQ have already taken the lead in the Agent field. However, the ultimate forms of AGI and even ASI will inevitably involve a deep integration of machine intelligence with the physical world—transitioning from &#34;data cognition&#34; to &#34;environmental interaction&#34; and &#34;action execution.&#34;
Ant Group understands that achieving this leap requires more than just the strength of a single company. It needs an ecosystem and a continuous influx of innovative blood. Over 70% of the participants come from 985 and 211 universities, including several students from prestigious overseas institutions. The drive and innovation embodied by these young people resemble Wang Xingxing's creation of the first robotic dog in his university dormitory.
As the competition concluded, the Wongtsai team from Zhejiang University won the championship with their outstanding autonomous control capabilities. However, for Ant Group, winning or losing is not the endpoint. Liu Yunhui predicts that it may take another five years or even longer for robots to truly enter households and provide nanny-level services. This is a long marathon.
Through the ATEC Elite Competition, Ant Group is waiting for the next &#34;Wang Xingxing&#34; who can break through real barriers with technology. In this extreme challenge of &#34;body&#34; and &#34;intelligence,&#34; the future of embodied intelligence may lie in the ability to stand up again after a fall

688688

Let AI venture into the physical world

ANT GROUP wants to bet more on "Wang Xingxing"