Embodied AI FRONTIER
- Your Embodied AI Intelligence Hub -
⌂ / Daily / ModelBest's Zou Zhensheng: Embodied AI Scaling Hinges on Reliability and Customer Economics
Industry

ModelBest's Zou Zhensheng: Embodied AI Scaling Hinges on Reliability and Customer Economics

At a Sept. 23 seminar in Beijing, the Digital Economy Research Center of China Economic Information Service hosted the "Xinhua Future Opportunity Dialogue: Breaking Through—From Demo-Level Intelligence to Scaled Application" high-quality development seminar on embodied AI. Zou Zhensheng, engineering vice president of Beijing ModelBest Intelligent Technology Co., Ltd., told attendees that embodied intelligence is not a one-company game but an ecosystem collaboration. Only by deepening the general model foundation can scenario transfer become engineering-meaningful. He stressed that the final mile of scaling is not flashy demos but reliability and making customers able to justify the economics.

He argued that the last mile of scaling embodied intelligence is not cool demos but reliability and economic viability for clients. That view framed his later discussion of data, scenarios and commercialization. He suggested the industry must move beyond presentation-level intelligence to large-scale application, where value is proven in operations rather than in showcases. He framed the challenge as moving from demonstrations that impress to systems that work reliably at scale.

On data-mix strategy, Zou said ModelBest was founded in 2022 and grew out of technology transfer from Tsinghua University's Natural Language Processing Lab. Different training stages require different data-mix strategies, a key guarantee for continuous model evolution. He explained that the right mix changes as training progresses, and that this discipline underpins long-term model improvement. He described the data strategy as a moving target rather than a fixed recipe.

He noted that human first-person-view data is currently the only embodied data supply that can scale to tens of millions of hours, giving it the greatest potential to serve as pretraining material for an embodied Scaling Law. But it lacks action labels and proprioception, requiring cross-embodiment alignment. Real-robot teleoperation data is currently the mainstay for post-training and offers the most direct success-rate gains on contact-intensive tasks, though collection costs are high. That combination defines the data challenge for scaling. The gap between available data and required labels remains a central engineering problem.

On the scenario ladder, Zou said the pace of embodied AI deployment depends on three dimensions: task definability—whether success criteria are clear; tolerance for failure costs; and difficulty of data feedback. These yield a clear sequence for commercialization. He grouped deployment scenarios into tiers based on those factors. This framework helps prioritize where embodied AI can deliver value first.

The first tier includes automotive and logistics sorting, where tasks are highly standardized and data feedback is smooth. The second tier is special scenarios, the third is commercial services, and the fourth is home scenarios. Home scenarios are highly non-standard, making task success hard to define uniformly and data feedback most difficult. That difficulty explains why home use is expected to lag other markets. By contrast, home environments demand broad generalization and tolerate little failure.

Discussing core constraints on downstream large-scale commercial use, he acknowledged that the real bottleneck is not only model capability but also reliability and systems engineering. The key is whether customers can make the economic math work. Reliability, not just intelligence, determines scalability, he said. Customers, he said, need clear returns before committing to deployment.

For the next two to three years, Zou predicted the edge large-model industry will unfold along a closed loop of perception, thinking and action capabilities. Language models are the most mature, multimodal models are next, and embodied VLA models are still evolving. The industry landscape will develop step by step rather than all at once. Each capability layer builds on the previous one.

In applications, automotive will scale first. Driven by rising model capability density and stronger edge-chip compute, consumer electronics will follow, and robots will become the largest commercialization scenario. Automotive is expected to lead, with consumer devices next and robotics offering the largest long-term market. That sequence reflects both technical readiness and market demand.

Zou called for embodied intelligence to be an ecosystem collaboration rather than a one-company game. He hoped more industry partners will join to build a multimodal collaborative data loop and evaluation system, while establishing publicly available dataset standards and trading mechanisms to break data silos. He also urged broader collaboration on shared evaluation and open data exchange. Such mechanisms, he said, are essential for sustainable progress.

✓ Verified 2026-09-25
Flash Briefs Live
▸ Today · Sep 25 Fri 3
▸ Yesterday · Sep 24 Thu 41
23:47
Industry 97
AI Demand Fuels High-End PCB Expansion as Robots Become Standard
21:34
Industry 98
The 19,999-Yuan Humanoid Robot: Who Buys First?
20:55
Industry 98
Changan Automobile Targets 2028 for Humanoid Robot Mass Production
14:30
Tech Agility 98
Agility CTO joins MIT CSAIL to discuss physical intelligence
14:00
Tech 98
Mexican EPICS Team Builds Portable Robotics Learning Platform
10:41
Industry 98
Robot Rental Demand Surges in China as Market Forecast to Grow Over 900%
10:04
Industry Astribot 98
Astribot to Showcase Real-World Robot Tasks at Pittsburgh Event
09:59
Industry Astribot 100
Astribot to Show Physical AI Stack in Pittsburgh, Sept 27-Oct 1
09:27
Industry 98
ANYbotics adds automated door access to ANYmal inspection robots
09:04
Industry 98
Amazon to Invest $100 Million in Indiana Manufacturing Facilities for Fulfillment and Robotics
08:30
Tech Sharpa 99
Sharpa to co-host ROCO and Origami challenges at IROS 2026
08:00
Industry Sharpa 100
Sharpa debuts D01 humanoid, W02 hand, AE01 glove at IROS 2026
07:52
Industry 98
Mifeng Technology Launches Crowdsourced Data Collection for Embodied AI Robots
07:08
Industry AGIBOT 100
AGIBOT's 20,000th Robot Rolls Off the Production Line
07:00
Industry lumos robotics 98
Lumos Robotics to Show NIX Bipedal Humanoid at IROS 2026 Booth
06:00
Industry LimX Dynamics 100
LimX Dynamics Opens Luna SDK for Full Development Pipeline
04:38
Industry DEEP Robotics 100
DEEP Robotics to Exhibit at IROS 2026 in Pittsburgh
04:04
Industry 98
Beta Infinite Launches First Robot, Positioned as a Life Teammate
03:57
Industry 98
Robots in Public Restrooms and Garment Factories Land Major Orders
03:55
Industry 98
Automakers Accelerate Humanoid Robot Push Despite Unprofitable EV Sales
03:48
Industry 98
Unitree Robots Walk Vogue World Runway Without Clothes, Spark Debate
03:40
Industry 99
Chunqiu Electronics Eyes Robotics as Key Expansion for Structural Components
03:33
Industry Unitree 100
Unitree: 19 humanoid robots dance with 120 dancers at WorldSkills
03:29
Tech 99
University of Washington's 1-Gram Jumping Robot Precisely Controls Jump Height
03:25
Tech 98
Polybot's Tomato Robot Learns to Harvest by Watching Humans
03:19
Industry 98
China Installs 354,000 Industrial Robots in 2025, Nearly 60% of Global Total
03:10
Capital 99
Grain Storage Robot Startup Xinhe Robotics Raises Tens of Millions in First Round
03:09
Capital 99
China's Weiyi Zhizao Secures First IPO Filing for Profitable Embodied AI Robot Maker
03:00
Capital 99
Dobot, Frontop Invest in Guangdong Embodied AI Firm as Capital Reaches 300M Yuan
02:55
Tech 95
Can GPT-6 Really Become a Robot Brain? Experts Debate World Models
02:25
Industry 95
Tesla Sources Chinese Parts for U.S.-Made Humanoid Robots
02:18
Tech 95
Xidian University and Intellifusion Launch Embodied Intelligence Joint Lab
02:13
Industry Galaxea Dynamics 96
Galaxea AI's R1 Pro robots deployed at Kunshan hospital
02:11
Industry 95
30x Stock Shangwei New Materials Enters Robotics, Shares Fall Three Days After Launch
02:09
Industry AGIBOT 95
AGIBOT's 20,000th robot rolls off the production line
02:06
Tech 95
Zheng Nanning: Embodied AI Is Not Humanoid Robots, Evaluation Isn't About Resembling Humans
01:57
Industry 95
Intel and Comtech Launch FoxJack X1, Fusing Robot Brain and Cerebellum on One Chip
01:51
Tech 97
Moying Technology Builds 'Android' for Industrial Embodied AI, Cuts Device Onboarding to 7 Days
01:31
Industry 95
AgiBot Delivers 20,000th Embodied AI Robot to Chimelong for Park Services
01:26
Tech 96
Deep Robotics Robot Dog Runs Full Course in Level-14 Typhoon with Zero Faults
00:50
Industry AGIBOT 96
AGIBOT Deploys 300+ Robots Across Chimelong Resort in Zhuhai