Xinhua News Agency's China Economic Information Service and CAICT jointly released the "2026 Embodied Intelligence Industry Innovation Report," proposing the XYZ curve and Worken

On October 29, 2026, at the AGIBOT INNOVATION DAY conference, Xinhua News Agency's China Economic Information Service, together with the Shanghai Industrial Innovation Center of the China Academy of Information and Communications Technology, officially released the "2026 Embodied Intelligence Industry Innovation Report."

The report is divided into five major chapters: Industry, Technology, Business, Ecosystem, and Outlook. It defines industry stages through the "XYZ Curve" of embodied intelligence industry development, drives technological evolution through "Three Intelligences in One," and measures productivity value through Worken, pragmatically advancing the innovative development of physical AI.

Pictured: The release of the "2026 Embodied Intelligence Industry Innovation Report" at the main forum of the AGIBOT INNOVATION DAY conference.

Artificial intelligence is accelerating its transition from the digital world into physical reality. Embodied intelligence has been systematically incorporated into the national future industry layout, and the track has ushered in 2026 as the inaugural year of the deployment era. The industry stands at a critical juncture in breaking through from technical validation to large-scale implementation, and productivity value is beginning to be unleashed at an accelerating pace. How to objectively view industry opportunities, identify real challenges, and clarify the true path forward for the industry—these are precisely the core concerns of the report. The report proposes that in 2026, embodied intelligence will begin moving from development demonstrations toward normalized operation in real scenarios, entering the deployment growth period. The yardstick for judging industry progress is also shifting from how many products were released and how many units rolled off the production line to how many robots truly enter factories, warehouses, and stores, continuously complete tasks, and deliver results.

As a practitioner and pioneer deeply involved in and continuously driving the industry's evolution, AGIBOT provided key research reference samples for the report. The report translates innovative practices from the industry front lines into industry consensus, recognizing both the long-term opportunities and real challenges contained in the track, while also clarifying more clearly the direction of the industry's forward evolution, driving the accelerated development of the embodied intelligence industry.

Reading through this report, one can grasp a main thread: use the "XYZ Curve" to judge where the industry has arrived, use "Three Intelligences in One" to see where technology is heading, and use "Effective Intelligent Work Hours (Worken)" to measure how much value there is—and all of this must ultimately be realized in the "deployment state": robots truly entering real scenarios, continuously working, and delivering results. 01 XYZ Curve: Identifying the Stage Coordinates of Industry Development

The report uses the "XYZ Curve" of embodied intelligence industry development to divide the industry into three major stages: development and early adoption, deployment growth, and deployment popularization. The progressive path from "embodied breakthrough" to "intelligence breakthrough" and then to "intelligence emergence" reflects the trends and laws of industry development. The "XYZ Curve" of embodied intelligence industry development was first proposed by AGIBOT in front-line industry practice. This curve is not merely a tool for dividing industry stages, but also an underlying logic system capable of unifying understanding across the entire industry. Starting from the main line of productivity implementation, it fully depicts the growth path of embodied intelligence from its beginnings to the coming years and establishes a common reference yardstick that is alignable, predictable, and capable of guiding investment decisions.

The "XYZ Curve" of Embodied Intelligence Industry Development

(Source: "2026 Embodied Intelligence Industry Innovation Report")

X Curve (2022–2025), the development and early adoption period, with the core being "embodied breakthrough." The industry conquered body motion capabilities, enabling robots to "move," with results concentrated in scientific research exploration and demonstrations; relying on China's complete supply chain, body capabilities accumulated rapidly. Y Curve (2026–2030), the deployment growth period, with the core being "intelligence breakthrough," and 2026 as the inaugural year of the deployment state. The industry's focus shifts from "whether it can move" to "whether it can work": robots leave exhibition halls and test grounds and enter real scenarios such as factories, logistics, and commercial cultural tourism for normalized operations. The deployment state is a comprehensive system capability: body, model, data closed loop, operations and maintenance, and business ecosystem—none can be missing. This is also the marker that "robots are no longer just demonstration demos."

Z Curve (after 2030), the deployment popularization period, with the core being "intelligence emergence." Large-scale continuous application of real machines accumulates massive interaction data, which, combined with model algorithm iteration, achieves scenario generalization and self-supervised evolution, unleashing trillion-level potential.

Standing in 2026, the industry is moving from the X Curve toward the Y Curve. Implementation in the deployment state has become the main line of the industry and is also the direction in which AGIBOT is firmly investing. 02 Three Intelligences in One: Grasping the Core Direction of Technological Evolution

The report points out the pain points of past technology paths: various capability modules were stacked and iterated separately, and prototype demonstrations were impressive, but in dynamic and ever-changing real environments, it was difficult to support stable, large-scale delivery. The report proposes that the industry is undergoing a paradigm upgrade: deeply coupling robot body hardware with the three major intelligences of motion, operation, and interaction, converging into an end-to-end collaborative "Three Intelligences in One" system. Among them, motion intelligence is the foundation, operation intelligence is the engine of labor productivity, and interaction intelligence is the entry point of service productivity; "in one" refers not only to a reliable body, but more fundamentally to the three major intelligences breaking free from fragmentation and moving toward end-to-end collaboration.

"Three Intelligences in One" is a development path with extremely high technical thresholds and dense challenges. Unlike polishing a single point capability in isolation, the four layers of capabilities are coupled with and constrain one another, imposing simultaneous iteration requirements on algorithm architecture, hardware body, perception devices, data systems, and systems engineering. R&D investment is large, and the technology implementation cycle is long, making it difficult to see visible results in the short term. However, from the long-term perspective of the industry, once the deep collaborative integration of the four layers of capabilities is completed and large-scale reuse across scenarios and tasks is achieved, technological accumulation will continue to produce compounding effects. This is the technology route with the highest compounding effect in the embodied intelligence field.

Having clearly identified the industry development stage and anchored the development path, a yardstick for measuring value is still needed to answer: how should the real output of robots be measured? This is precisely the purpose of the report's introduction of "Effective Intelligent Work Hours." 03 Worken Effective Intelligent Work Hours: Establishing a New Yardstick for Physical AI Productivity

Digital AI measures information processing in Tokens, while physical AI needs a new yardstick for measuring physical labor. The report introduces "Effective Intelligent Work Hours" (Worken)—originally proposed by AgiBot and modeled on three factors: the number of active robots, autonomous working hours, and output efficiency ratio—to define the true output value of embodied intelligence.

Overview of the "Effective Intelligent Work Hours" indicator

(Source: 《2026 Embodied Intelligence Industry Innovation Report》)

Offline volume represents manufacturing capability, shipment volume represents contracted delivery, and only deployment volume and Effective Intelligent Work Hours represent the real landing of productivity—only when autonomous tasks are continuously completed in real scenarios are Effective Intelligent Work Hours generated. Traditional hardware begins depreciating immediately after delivery; embodied intelligence products with a data closed loop, by contrast, continuously iterate in real applications—the more they are deployed, the smarter they become. The report also records AgiBot's original Worken sample library, turning abstract robot value into visible, calculable operational results and providing vivid evidence of large-scale deployment. Worken provides a unified yardstick for measuring labor value and commercial settlement, guiding the industry to break away from parameter involution and return to real output, and driving embodied intelligence to complete the historic leap "from quantitative change to qualitative change."

AgiBot Worken Original Sample Library

(Source: Main Forum of the AgiBot Innovation Day Conference)

Coordinates, paths, and yardsticks must ultimately all face the same testing ground—the deployment state: whether the technology can withstand long-term testing in real scenarios, and whether value can continuously accumulate in real applications. 04 Anchoring the development direction, industrial practice rooted in the deployment state

The report points out that physical AI cannot be achieved through concepts and demonstrations; it needs continuous refinement in real scenarios, and uses the "scenario complexity × task complexity" coordinate to stratify and position deployment scenarios: starting from simple tasks in simple environments to validate economics, then advancing step by step toward complex tasks in complex environments. In terms of rhythm, the short term prioritizes high-certainty industrial scenarios such as handling, sorting, and inspection; the medium term penetrates commercial services and complex industry; the long term moves toward all scenarios. Each preceding scenario accumulates data and experience for the next, and scenario expansion is a snowballing progression rather than a leap, advancing from simple to complex in echelon fashion. Use the XYZ curve to see the stage clearly, the integration of the three intelligences to anchor the path, and Worken to measure real output—the development of physical AI lies in pragmatism. Only when more and more robots truly take up posts and continuously produce "Effective Intelligent Work Hours" will the era of physical AI truly arrive—the report judges that the emergence of physical AI will reach its endgame through five key leaps: stable and usable bodies, vertical-domain large-scale deployment, generalized embodied foundation models, collective collaborative evolution, and intelligent self-lifting. Whether a full-stack closed loop can be achieved is the core variable for the industry to pass through the Y curve and move toward the Z curve, and it also confirms AgiBot's route choice of full-stack R&D rooted in the deployment state.

Schematic diagram of the five leaps (Source: Main Forum of the AgiBot Innovation Day Conference)

The deeper value of the report lies in providing a set of industry-consensus language and jointly directing it toward the value creation of real productivity. The value focus of the report is also very clear: relying on a systematic analytical framework combining theory and practice, it provides reference coordinates for industry practitioners to grasp the pace of development, for investment decision-makers to identify real value, and for policymakers to calibrate the direction of policy measures. As the concluding part of the report points out: the ultimate competition in embodied intelligence is no longer merely a race of single-point technologies among enterprises, but a systemic contest of software-hardware collaboration, engineering implementation, industrial organization methods, and governance capabilities. AgiBot is willing to join all parties in being steadfast pragmatists of this era.

Source: AgiBot official WeChat account, original article "Understand in One Article" Download the report

Official WeChat account "Smart City Industry Analysis": reply in the background with "2026 Embodied Intelligence Industry Innovation Report" or "Embodied Intelligence Report" to obtain the report download address

www.smartcity.team

Disclaimer: Please indicate the source as "Smart City Industry Analysis (www.smartcity.team)" and the relevant author when reprinting. Except for original content, some information on this site comes from the Internet. If the relevant content involves infringement of rights, please contact the contact information at the bottom to delete it.

[Digital Technology Special Research] Community

In-depth peer exchange, focused on smart city and digital technology industry research and material accumulation.

Continuously updating the latest industry white papers, industry research, policy developments, and implementation solutions to help industry practitioners efficiently accumulate knowledge and share experience.

Massive premium content continuously iterated, providing frontier perspectives, connecting industry talent, carrying out case co-creation and resource exchange, and jointly building a high-quality industry exchange atmosphere.

#post-##

Recommended for you

Jiaziguangnian Think Tank "2026 Agent Commercialization and Competition Path Research Report": analysis from five aspects: product and demand evolution, commercial value formation, market ecosystem and enterprise deployment, application landscape and commercialization practice, and industry evolution and competition outlook

"Embodied Intelligence Development Report (2026)": embodied intelligence is at a transitional node toward pilot applications; the industry is heating up while revealing a chill, the technical system composed of the four elements of models, data, bodies, and computing power is evolving collaboratively, and industrial chain division of labor is refining, but a sustainable commercial closed loop still needs to be established

National Public Data Operation Development Report (2025-2026): there are 76 public data resource authorized operation platforms or zones that have been publicly disclosed as operating or under construction nationwide, with 202 implementation institutions and 224 operating institutions identified

Embodied Data Development Report 2026: detailed explanation of the development status, meaning and classification, industrial chain and business models, distinctive cases, and development trends of the embodied data industry

Device-side Intelligence Industry Development Report (2026): systematically reviews the overview, technical system, product forms, and application scenarios of the device-side intelligence industry

"Green Token Factory Technical White Paper": systematically elaborates the five-layer technical system of the green token factory: "energy infrastructure—computing infrastructure—Token production—Token operation—Token circulation"