As a pivotal element in the transition of artificial intelligence from “disembodied cognition” to “embodied cognition,” embodied intelligence has gained significant attention, particularly with its inclusion in the 2025 Government Work Report as a key future industry. This recognition underscores its role alongside biomanufacturing, quantum technology, and 6G in shaping national technological innovation strategies. Embodied intelligent robots, which integrate AI into physical entities to enable perception, learning, and environmental interaction, are emerging as strategic industries capable of reshaping global manufacturing competitiveness. In Suzhou, a major hub in the Yangtze River Delta, the development of the embodied intelligent robot industry has shown remarkable progress, with over 600 related enterprises and an industrial scale reaching 139.5 billion yuan in 2024. However, as an emerging field, challenges persist in achieving the goals outlined in the “Suzhou Embodied Intelligent Robot Industry Innovation and Development Three-Year Action Plan (2025–2027),” such as establishing Suzhou as a national source of technological innovation, a high-end manufacturing cluster, and a model city for demonstration applications. This study aims to address these challenges by constructing a comprehensive evaluation system to assess the development level and potential of core enterprises in the embodied intelligent robot sector, identifying bottlenecks, and proposing targeted strategies for high-quality growth.
The theoretical foundation of embodied intelligent robots traces back to Alan Turing’s early propositions in the 1950s, emphasizing machines’ autonomy in interacting with environments, perceiving, decision-making, planning, and executing tasks. Embodied intelligent robots are characterized by their physical presence, ability to sense environments, comprehend tasks, and autonomously perform actions. Unlike traditional industrial robots, they exhibit superior adaptability, task understanding, and decision-making capabilities. Core features include: (1) advanced perceptual abilities through multi-modal sensors like vision, touch, and hearing; (2) robust cognitive capacities for interpreting complex instructions and reasoning; (3) flexible execution skills in unstructured environments; and (4) learning and evolution capabilities through experience accumulation. These intelligent robots manifest in various forms, such as fixed-base robots, humanoid robots, mobile robots, and bionic robots, each tailored to specific applications like industrial automation, healthcare, and environmental monitoring.
Enterprises in the embodied intelligent robot domain are defined as high-tech organizations engaged in fundamental research, core technology development, key component manufacturing, and system integration. To qualify, these entities must meet criteria including independent legal status, substantial R&D investment (e.g., at least 5% of revenue), core patents, product commercialization in sectors like industry or healthcare, and compliance with national and local policies. The industry encompasses an integrated chain from upstream components (e.g., sensors, controllers, and AI algorithms) to midstream robot manufacturing and downstream application services. This ecosystem is driven by technological innovation and policy support, facilitating expansion into diverse fields such as smart manufacturing, commercial services, and public welfare.
Suzhou’s development in the embodied intelligent robot industry is built on a strong manufacturing base and a favorable business environment. The city ranks among the top in China’s robot city comprehensive strength assessments, hosting nearly 800 enterprises in robotics and key components, including 24 national-level “little giant” specialized firms and 14 listed companies. The workforce comprises approximately 63,000 professionals, and in 2024, the industry generated 3,125 patent grants, a 18.7% year-on-year increase, with inventions accounting for 42%. Leading companies like Ecovacs and Harmonic Drive have achieved global milestones, such as shipping over 10 million service robots and producing more than 2 million precision reducers annually. Innovation platforms, including the Suzhou Embodied Intelligent Robot Comprehensive Innovation Center, have attracted 20 high-level R&D institutions, undertaken 83 provincial-level projects, and secured over 800 million yuan in government funding. Additionally, 136 “robot+” demonstration projects have been implemented, spanning areas like smart factories and urban inspection, catalyzing upstream and downstream investments of around 21 billion yuan.
Regional collaboration plays a crucial role in Suzhou’s strategy, with integration into the Yangtze River Delta initiative fostering synergies with Shanghai and Hangzhou. For instance, partnerships with Shanghai’s Zhangjiang AI Island and Hangzhou’s Future Sci-Tech City focus on joint R&D in core components like reducers and servo systems. Internationally, Suzhou has established links with global innovation networks, attracting multinational corporations to set up R&D centers and hosting forums to pool top-tier resources. Policy measures, such as the “Several Measures to Support the Innovation and Development of the Embodied Intelligent Robot Industry in Suzhou,” provide financial incentives, including up to 200 million yuan for national key laboratories and 20 million yuan for provincial labs, aiming to overcome technological bottlenecks and accelerate industrialization.
To scientifically evaluate the high-quality development of Suzhou’s embodied intelligent robot enterprises, this study employs a questionnaire survey method, analyzing responses from 42 valid samples collected from key industrial parks. The evaluation system is structured into target, criterion, and indicator layers, focusing on four dimensions: technological R&D capability, technological innovation level, commercialization efficiency, and international influence. A Likert 5-point scale was used to measure the importance of each indicator, and the data underwent reliability and validity tests, principal component analysis (PCA), and entropy weight method to ensure robustness.
The reliability test yielded a Cronbach’s Alpha coefficient of 0.926, indicating high internal consistency. The KMO measure of sampling adequacy was 0.816, and Bartlett’s test of sphericity showed a significance level of 0.000, confirming the suitability for factor analysis. PCA extracted four principal components with eigenvalues greater than 1, explaining 65.30% of the total variance. After rotation using the varimax method, 12 key indicators were retained, as shown in Table 1.
| Indicator | Technological R&D Capability | Technological Innovation Level | Commercialization Efficiency | International Influence |
|---|---|---|---|---|
| Originality of Fundamental Algorithms | 0.83 | 0.18 | 0.12 | 0.09 |
| Self-Sufficiency Rate of Key Components | 0.79 | 0.15 | 0.21 | 0.14 |
| Scale of High-Quality Open Datasets | 0.77 | 0.22 | 0.19 | 0.11 |
| Source Technology Supply | 0.16 | 0.81 | 0.14 | 0.20 |
| Proportion of High-Value Patents | 0.19 | 0.78 | 0.17 | 0.10 |
| Enterprise Incubation Success Rate | 0.21 | 0.75 | 0.12 | 0.15 |
| Technology Maturity Transition Cycle | 0.15 | 0.18 | 0.82 | 0.14 |
| Density of Scenario Deployment | 0.11 | 0.21 | 0.80 | 0.19 |
| Proportion of Commercialization Revenue | 0.14 | 0.16 | 0.76 | 0.22 |
| Proportion of PCT Patents | 0.12 | 0.14 | 0.18 | 0.84 |
| Number of Overseas Joint Laboratories | 0.19 | 0.15 | 0.17 | 0.80 |
| Contribution to International Standards | 0.10 | 0.18 | 0.20 | 0.77 |
The four principal components were named as follows: (1) Technological R&D Capability, reflecting core technical prowess and autonomy; (2) Technological Innovation Level, indicating front-end innovation sourcing and quality; (3) Commercialization Efficiency, measuring the speed and breadth of market application; and (4) International Influence, representing global impact and collaboration. The entropy weight method was applied to determine the relative importance of each indicator within these components. The entropy value $$ e_j $$ for each indicator is calculated as:
$$ e_j = -k \sum_{i=1}^{n} p_{ij} \ln p_{ij} $$
where $$ p_{ij} = \frac{x_{ij}}{\sum_{i=1}^{n} x_{ij}} $$, $$ k = \frac{1}{\ln n} $$, and $$ n $$ is the number of samples. The entropy weight $$ w_j $$ is then derived as:
$$ w_j = \frac{1 – e_j}{\sum_{j=1}^{m} (1 – e_j)} $$
where $$ m $$ is the number of indicators. The results, presented in Table 2, show that self-sufficiency rate of key components and originality of fundamental algorithms have the highest weights, emphasizing their critical role in the development of embodied intelligent robots.
| Criterion Layer (Principal Component) | Indicator Layer | Information Entropy | Entropy Weight | Weight Ranking |
|---|---|---|---|---|
| Technological R&D Capability | Self-Sufficiency Rate of Key Components | 0.698 | 0.201 | 1 |
| Originality of Fundamental Algorithms | 0.713 | 0.187 | 2 | |
| Scale of High-Quality Open Datasets | 0.735 | 0.164 | 4 | |
| Technological Innovation Level | Source Technology Supply | 0.721 | 0.179 | 3 |
| Proportion of High-Value Patents | 0.754 | 0.146 | 5 | |
| Enterprise Incubation Success Rate | 0.772 | 0.128 | 7 | |
| Commercialization Efficiency | Density of Scenario Deployment | 0.764 | 0.136 | 6 |
| Technology Maturity Transition Cycle | 0.789 | 0.111 | 8 | |
| Proportion of Commercialization Revenue | 0.805 | 0.095 | 9 | |
| International Influence | Contribution to International Standards | 0.819 | 0.081 | 10 |
| Number of Overseas Joint Laboratories | 0.831 | 0.069 | 11 | |
| Proportion of PCT Patents | 0.846 | 0.054 | 12 |
Based on the evaluation results, a four-pronged strategy system is proposed to foster the high-quality development of Suzhou’s embodied intelligent robot industry: Core Foundation Building, Innovation Engine Driving, Commercialization Acceleration, and Global Leapfrogging.
First, in Core Foundation Building, Suzhou should establish a specialized fund for the localization of key components, such as reducers, servo motors, and tactile sensors, using mechanisms like “unveiling the list and appointing the best” to invite global teams to tackle technological bottlenecks. This fund would provide equity investments and R&D grants to reduce market entry costs. Concurrently, an open data lake and supercomputing platform for embodied intelligent robot training should be developed, aggregating real-world data from Suzhou and the Yangtze River Delta. Enterprises contributing to this data asset could receive computing vouchers, fostering a closed-loop ecosystem of data, algorithms, and hardware. This approach aims to enhance the self-sufficiency rate of key components and originality of fundamental algorithms, which are vital for the intelligent robot sector’s resilience and innovation.
Second, for Innovation Engine Driving, Suzhou ought to create “R&D enclave + pilot acceleration” models by collaborating with top universities like Fudan University and Zhejiang University to set up R&D centers in academic hubs. These enclaves would attract teams with groundbreaking projects, offering long-term funding and talent incentives. Within Suzhou, pilot testing bases with multi-modal robot testing and simulation capabilities should be expanded to expedite the transition from prototypes to mass production. Additionally, “high-value patent cultivation vouchers” could reward enterprises for international patent applications and citations, strengthening intellectual property barriers. This strategy targets the enhancement of source technology supply and high-value patent proportions, crucial for sustaining innovation in the intelligent robot industry.
Third, Commercialization Acceleration can be achieved through the “Thousand Enterprises, Ten Thousand Robots” demonstration project, which releases annual opportunity lists for applications in smart healthcare, manufacturing, and agriculture. A “scenario + order” joint bidding model would encourage enterprises to partner with scenario providers, with government subsidies and risk-sharing mechanisms for successful demonstrations. Furthermore, a government-guided fund tied to commercialization revenue could co-invest with social capital, directly linking scene orders to enterprise income and valuation. This focuses on improving the density of scenario deployment and reducing the technology maturity transition cycle, thereby speeding up the market penetration of intelligent robots.
Fourth, for Global Leapfrogging, Suzhou should leverage free trade zone policies to implement “zero-tariff import of equipment for international joint laboratories,” simplifying cross-border processes for advanced research tools. Streamlined services for foreign high-end talents, including visa and work permit integration, would attract global experts to establish joint labs. A “International Standards Navigation Reward Plan” could incentivize enterprises to lead in setting ISO/IEC standards for embodied intelligent robots. Regular participation in global exhibitions, such as Germany’s Hannover Messe or Japan’s iREX, would showcase Suzhou’s models, enhancing its international influence. This strategy aims to boost the proportion of PCT patents and overseas joint laboratories, positioning Suzhou as a global leader in the intelligent robot arena.
In conclusion, the embodied intelligent robot industry represents a transformative force in AI and manufacturing, with Suzhou poised to capitalize on its strengths through a structured evaluation system and targeted strategies. By emphasizing core technological foundations, innovation drivers, efficient commercialization, and global integration, Suzhou can overcome existing bottlenecks and achieve high-quality development. This approach not only aligns with national strategic goals but also fosters a sustainable ecosystem for intelligent robots, contributing to regional and global technological advancement. Future efforts should focus on continuous monitoring and adaptation of these strategies to evolving market and technological trends.