Physical AI OS & World's First Industrial Embodied 6S Store Debut at Forum

On October 12, the "Industrial Embodied Intelligence: Manufacturing Without Boundaries" 2026 Industrial Embodied Intelligence Industry Innovation Summit was held at the National Exhibition and Convention Center as a parallel event of the 26th China International Industry Fair. Over 300 representatives from government, research institutes, industry, and finance gathered to discuss technical breakthroughs and large-scale deployment paths for embodied intelligence in industrial scenarios.

Physical AI Industrial Embodied Operating System: Enabling Multi-Robot Collaboration in Digital Factories

A core bottleneck in industrial embodied intelligence is scarce real production-line data, high validation costs, and the "last mile" gap between simulation and real machines. To address this, Lingqiao Intelligence and Songying Technology announced an ecosystem partnership and jointly released the Physical AI Industrial Embodied Operating System and the Lingqiao Manipulation Module.

The system is built on Songying Technology's self-developed ORCA OS, China's first physical AI operating system for multi-morphology robot collaborative training. ORCA OS enables heterogeneous agents — humanoid robots, quadrupeds, drones — to train collaboratively and execute tasks in the same digital factory scene, closing the full R&D loop from digital training to real-machine deployment. Its core capabilities include:

Scene construction

Synthetic data generation

Multi-physics fusion simulation

Parallel accelerated training

SIL/HIL dual-loop verification

Sim2Real real-machine deployment

These integrate previously fragmented toolchains into a unified workflow that is observable, developable, verifiable, and deployable. On the compute side, ORCA OS adopts a hardware-decoupled architecture and has been adapted to international vendors (NVIDIA, AMD, Intel) and domestic mainstream GPUs (Moore Threads, Iluvatar CoreX, Enflame, Hannbo Semiconductor, Lisuan Technology), making it the industry's only physical AI simulation platform compatible with multiple hardware architectures.

Lingqiao Intelligence leads the dexterous manipulation module, leveraging its full-stack capabilities in manipulation models, data engines, and dexterous hardware to convert manipulation feel into robot-learnable, practicable, and transferable skills. The partnership completes the closed loop from data generation, model training, simulation verification to real-machine deployment, and opens the ecosystem so that robotic arms, quadrupeds, and humanoids can learn to collaborate in the same physical world and flow into production processes.

Also unveiled was the new 021B — a 21-degree-of-freedom dexterous hand with "brain" and "cerebellum" embedded in the palm: the brain handles task understanding and planning decisions, while the cerebellum handles fine motor control and real-time adjustment.

Three-Layer Architecture Driving 13-Company Ecosystem Co-Construction

Lingqiao Intelligence released a three-layer industrial embodied intelligence architecture:

Base layer: compute, data, pipelines, algorithm models, simulation systems — the technical foundation.

Body layer: robotic arms, quadruped robots, humanoid robots, and other hardware platforms.

End-effector layer: high-performance end-effector development, giving industrial robots core precision operation capabilities.

Above the three layers lie full-chain "R&D, production, supply, sales, service" application scenarios. As the core lead of the end-effector layer, Lingqiao Intelligence unites base-layer, body-layer, and application-scene partners to push industrial embodied intelligence onto real production lines.

At the forum, Lingqiao Intelligence, together with the Ministry of Industry and Information Technology's Key Laboratory of Robot Quality Testing and Evaluation, Songying Technology, Tianji Intelligence, Ruierman, Jaka, and nine other companies/institutions (13 total), launched the Industrial Embodied Intelligence Industry Partner Program. It covers key links such as dexterous hands, robotic arms, quadrupeds, simulation platforms, and testing/certification, aiming to accelerate technology transfer to physical lines and build a complete ecological network from core components to whole machines to scenario applications.

World's First Industrial Embodied Intelligence 6S Store: Transaction Infrastructure Connecting Supply and Demand

If the physical AI OS solves "can the technology work?", the 6S store answers "who buys the product and where to buy the solution?" Despite maturing technology bases and robot platforms, a wall remains between supply and demand: suppliers don't know who to sell to, buyers don't know where to find solutions, and hardware, software, and agents are scattered across industries, unable to be procured like standard industrial equipment.

To bridge this, Shanghai AI Research Institute, Shanghai Electric, Lingqiao Intelligence, and JD Retail jointly launched the world's first "Industrial Embodied Intelligence 6S Store," operated by Lingji Zhilian (a joint venture of Shanghai Electric and Lingqiao Intelligence). The platform is a market-oriented transaction service platform for both sides of industrial embodied intelligence: one end connects hardware vendors, software providers, and agent developers for centralized showcase of cutting-edge solutions; the other end connects manufacturers and integrators, offering "one-stop" services from scenario publishing, product selection, to solution procurement.

The 6S store's core lies in leveraging JD Retail's supply chain system and transaction capabilities — not just traffic entry, but a mature merchant onboarding system, massive enterprise procurement traffic, and complete supply chain fulfillment from order to delivery. This base enables rapid onboarding of more merchants and service to more buyers, shifting the commercial model from scattered single transactions to standardized, long-term ecological operations. For industry, it's not about who has the longest plank, but about building a true "hexagonal warrior" that tightens every screw on the production floor. Only with wide entry and strict exit can embodied intelligence new technologies truly integrate with industry and drive next-generation intelligent manufacturing.

Ecosystem Co-Construction and Industry Research Advancing in Parallel

The forum also released multiple industry collaboration and ecosystem achievements:

Ecosystem co-construction: Guided by the "Global Industrial AI Guidelines," Shanghai AI Research Institute, Global Industrial AI Alliance Excellence Center, Schneider Electric, Siemens, Shanghai Electric, Bank of Shanghai, China Unicom, SGS, and others launched the "Global Industrial Embodied Intelligence Partner Ecosystem Co-Construction Plan," advancing five key areas: standard formulation, evaluation system construction, case compilation, solution library collection, and ecosystem building.

Carrier platform: The Industrial Embodied Intelligence Carrier Platform, jointly built by Minhang District Government and Shanghai Electric, was unveiled. Relying on real industrial production lines, it provides real-scene training spaces that are trainable, testable, and verifiable, outputting the "Shanghai Path" for industrial embodied intelligence nationwide.

Industry research: Shanghai AI Research Institute, together with Shanghai Modern Service Industry Federation, Yicai, Huizheng Finance, Donghao Lansheng, Kingdee, and Shanghai Yangtze River Delta Advanced Manufacturing Development Research Institute, released the "2026 AI + High-End Manufacturing Industry Application Map," focusing on AI + industrial software/platforms, AI + intelligent manufacturing equipment, AI + industry-specific equipment, providing cross-industry, comparable, and implementable "Shanghai Path" for AI industry integration policy implementation. The Shanghai AI Research Institute's Industrial AI Center then gave a professional interpretation of core AI application scenarios and implementation paths in industry.

In the keynote session, Chinese Academy of Sciences Academician Ding Han, China Electronic Information Industry Development Research Institute Technology Division Director Dong Kai, Schneider Electric China General Industry Director Jin Shi, and National-Local Co-Built Humanoid Robot Innovation Center Technology Director Xing Boyang shared insights from basic research, industrial policy, industry practice, and innovation platform construction. A roundtable hosted by Lingqiao Intelligence Government Affairs and Ecology Cooperation Director Zhang Long featured China Mobile Shanghai Industrial Research Institute Chief Engineer Luo Chenchi, Yunshenchu Ecology Cooperation Director Cheng Yuxing, JD Retail Intelligent Robot Business Unit Senior Manager Yu Yibiao, Zhenjue Technology Founder & CEO Liu Qi, and Sazhi Intelligence Chairman & CEO Zhang Jianzheng, discussing "Breaking Through Scenario Deployment, Building Industrial Embodied Ecological Systems."

Centered on "Embodied Intelligence Landing, Industry First," the summit convened multi-stakeholder forces from government, industry, academia, research, and finance, building a collaborative platform for "policy — technology — scenario — ecology." As co-organizer and deep participant, Lingqiao Intelligence showcased a complete industrial landing path from dexterous manipulation core technology, physical AI industrial embodied OS, three-layer industrial architecture, to 6S store transaction platform. Going forward, all parties will rely on the industrial embodied intelligence carrier platform, global partner ecosystem co-construction plan, and industrial landing framework to advance the "real-scene training — data precipitation — product iteration — large-scale deployment" closed loop, contributing industrial strength to promoting new industrialization and developing new quality productive forces.