On March 26, Nanjing officially unveiled the Robot Industry Tackling and Promotion Office. With institutional innovation as a breakthrough point, the office is fully committed to fostering more competitive innovative enterprises and racing to climb new heights in the development of the robot industry.

Amid Nanjing’s booming robot industry cluster, Nanjing Estun Kuzhuo Technology Co., Ltd. (hereinafter referred to as "Estun Kuzhuo") shines like a dazzling new star. It has a "distinguished lineage"—incubated with key support from Estun Automation Group (hereinafter referred to as "Estun"), a leading domestic industrial robot enterprise; it has made a strong "breakthrough"—successfully developing a complete humanoid robot product in just over two years; and it shoulders a "heavy responsibility"—as the first direct investment project of the first batch of industrial special funds under Jiangsu Province’s Strategic Emerging Industry Master Fund (hereinafter referred to as "Provincial Master Fund").

In the second quarter of this year, Estun Kuzhuo’s second-generation humanoid robot with independent walking function—Codroid 02—will be launched. At a time when technological breakthroughs and commercial implementation are accelerating, how will Estun Kuzhuo, with the pragmatic genes and "hardcore" strategies of the "industrial sector", embrace the era of transformation that is approaching at full speed?

Stepping into Estun’s product exhibition hall, one is immediately greeted by Kuzhuo’s humanoid robot CODROID 01. This robot attracted widespread attention from the industry at the 24th China International Industry Fair held in September last year.

CODROID 01 stands 170 cm tall, with a total weight of less than 80 kg, and features a sleek gray-black mecha-style design. It integrates 44 active degrees of freedom throughout its body, achieving full-scale design from the head to the hands, arms, legs, waist, and ankles.

What upgrades will the upcoming Codroid 02 bring? Li Yuanping, Co-founder and General Manager of Estun Kuzhuo, revealed: "The second-generation product has achieved significant optimizations in both hardware structure and software architecture. The intuitive improvements include more stable gait, more flexible fingers, while also reducing weight and costs."

It is reported that Codroid 02 will demonstrate independent walking capability for the first time. Its total weight will be reduced by more than 10 kg, yet the strength and dexterity of its arms will be greatly enhanced. Compared with the first-generation product, which could only demonstrate basic action combinations, the second-generation robot has acquired intelligent decision-making capabilities and can perform simple handling tasks. The load capacity of its dual arms exceeds 10 kg, and the overall cost of the robot has been further reduced.

Founded in July 2022, Estun Kuzhuo started with single-arm collaborative robots, then tackled the R&D of mobile composite robots with integrated sensing and control of hands, feet, eyes, and "brain", and finally achieved the production of complete humanoid robots. In just over two years, it has completed the technical "climb" that takes many of its peers years to accomplish.

The "Kuzhuo Speed" stems from its unique industrial genes. Incubated independently by Estun—a leading domestic industrial robot enterprise—Estun Kuzhuo, from its establishment, has targeted the cutting-edge field of "embodied intelligence"—where robots can perceive the external environment through sensors and independently make decisions and execute corresponding actions based on changes in the environment.

"Embodied intelligence will be the core battlefield for the next generation of robots, and humanoid robots are its ideal carrier," Li Yuanping stated frankly.

In response, Estun assembled Kuzhuo as an "elite team". Its members possess profound technical accumulation in the field of embodied intelligence, and Li Yuanping himself is a core member of the "unveil the list and take command" R&D project for humanoid robots under the Ministry of Industry and Information Technology.

At the same time, relying on Estun’s accumulation in motion control, mechanical bodies, and industrial scenario data, Estun Kuzhuo has built rapid productization capabilities. Estun’s supply chain and customer resources have also provided Kuzhuo with ready-made commercial channels, enabling it to quickly verify technologies in industrial scenarios and emerge as a "dark horse" in the humanoid robot track.

Entering 2025, the humanoid robot market has become "boiling with enthusiasm": humanoid robots have become "standout stars" by dancing and doing somersaults, fueling public enthusiasm to a climax; startups are competing to lay out their plans, cross-industry giants are rushing to enter the field, and a "hundred-robot battle" is on the verge of breaking out; technological evolution is "advancing by leaps and bounds", and the commercialization process has accelerated significantly...

Amid this fierce global race, Estun Kuzhuo has demonstrated the unique stance of the "industrial sector"—not chasing the spotlight, but focusing on tackling "hard bones", with its own strategy of "advancement and retreat".

Considering that embodied intelligence is still in the embryonic stage of technological development, while consumer applications are lively and noisy, the specific scenarios and tasks are more complex. Estun Kuzhuo has firmly committed itself to rigid-demand fields such as industrial manufacturing and warehousing logistics, focusing on high-certainty tracks. The company is confident: "Humanoid robots will first be deployed in these labor-intensive scenarios, and the potential market scale is huge."

Furthermore, technological breakthroughs in humanoid robots focus on two core areas: the "brain", which focuses on cognition and decision-making, and the "small brain" (motor control system), which is related to motion control and determines the robot’s physical execution capabilities. Estun Kuzhuo established its technical route at an early stage: for the "brain", it focuses on model simplification and fine-tuning rather than developing large models independently; for the "small brain", it must independently overcome the core technologies of the robot’s body—only by solidifying underlying technologies such as motion control can it quickly integrate with the evolved "brain" and achieve an intelligent leap.

This strategy was quickly verified. At the beginning of 2025, the technological innovation wave triggered by DeepSeek swept across many industries. After equipping with large models such as DeepSeek, the "brain" capabilities of humanoid robots have advanced by leaps and bounds, becoming more similar to humans in interaction, understanding, and decision-making. However, the industry has also revealed a technical imbalance of "strong cognition but weak execution"—and this is exactly the "battlefield" where Estun Kuzhuo has been accumulating strength for a long time.

"Data is the ‘soul’ of humanoid robots, but there is a world of difference in the difficulty of obtaining data for the ‘brain’ and the ‘small brain’," Li Yuanping said.

Compared with the training data for large models, which comes from massive text, voice, images, etc. on the Internet and can be collected in batches at low cost, collecting data for the "small brain"—whether it is simulation-generated data or real-machine data—is far more difficult than collecting data for large models.

For example, to let a robot "recognize a water cup" only requires a hundred images, but to let it "hold the water cup steadily" not only requires it to actually hold the water cup a hundred times, but also requires continuous "training" based on data such as the tilt angle of the cup, the grip strength of the fingers, and the shaking of the water flow.

A highly flexible robotic hand requires more than 20 degrees of freedom and integrates nearly a hundred sensitive contacts for vision, touch, etc. In Li Yuanping’s view, fitting so many components into "one hand" involves a level of technical complexity far exceeding that of robot limb R&D. Currently, Estun Kuzhuo mainly starts with basic scenarios such as laboratories and unordered grasping to train the robot’s hand-eye coordination capabilities. At the same time, it collects multi-modal data and builds a data management platform to improve the humanoid robot’s operational capabilities in different scenarios.

"Commercialization does not need to wait for perfect technology; good industrial products are refined through customer use," he said. At this stage, limited intelligence already has commercial value in vertical scenarios. Only through continuous guidance from customers and scenarios can humanoid robots gradually move from "being able to think" to "being able to work".

Just as a humanoid robot requires precise coordination of all its joints to walk steadily, industrial leapfrogging also requires the "multi-driven" synergy of technology, capital, and policies.

On November 11 last year, Estun Kuzhuo became the first direct investment project of the first batch of industrial special funds under the Provincial Master Fund. "This financing marks Estun Kuzhuo’s first introduction of external investors. The financing window was short, and the capital gap was large," a relevant person in charge of the First Investment Department of Nanjing Innovation Investment Group (hereinafter referred to as "Nanjing Venture Capital Group") revealed. To give full play to the role of government-guided investment funds in boosting and guiding the cultivation of new quality productive forces, on the one hand, the Jiangsu Nanjing Software and Information Service Industry Special Master Fund, managed by Nanjing Venture Capital Group, directly invested tens of millions of yuan in Estun Kuzhuo; on the other hand, through the active communication and coordination of Nanjing Venture Capital Group, the Estun Kuzhuo project gained the favor of the National Advanced Manufacturing Fund and received an investment of 100 million yuan, greatly alleviating the "capital hunger" in the startup stage.

It is understood that the Provincial Master Fund adheres to the positioning of "patient capital" and does not take profit as its main goal. The aforementioned person in charge said, "At present, the efficiency, cost, and intelligence level of embodied intelligent robot products still need to be optimized. However, industry customers have application scenarios, and Kuzhuo has solutions. We look forward to the company presenting a ‘hardcore answer’ in the field of embodied intelligent robots for Nanjing."

It is worth mentioning that on March 26, Nanjing unveiled four Industry Tackling and Promotion Offices, with the robot industry being one of them. The Industry Tackling Office has set up teams for industrial investment promotion, project promotion, achievement transformation, and enterprise cultivation and services. At the same time, it integrates the strengths of leading chain enterprises, universities, research institutes, industry organizations, and investment institutions.

The industry regards 2025 as the "first year of commercial mass production" of humanoid robots. Amid the wave of "a hundred robots competing for development", Li Yuanping frankly stated: "The industry has entered a stage of separating the genuine from the fake. Whether it is the company becoming the first ‘provincial strategic emerging industry direct investment project’ or the Nanjing municipal government’s full efforts in industrial tackling, both will install a powerful ‘power system’ for the company’s development and help the enterprise realize the beautiful vision of ‘human-machine integration’."

For this "positioning battle" in the humanoid robot track, Estun Kuzhuo has formulated a 2025 technological tackling roadmap: in terms of hardware breakthroughs, it will launch an industrial-grade humanoid robot with a single-arm load of 20 kg in the middle of the year, with a dexterous hand featuring higher degrees of freedom; in terms of software iteration, it will release the second-generation software control system in the third quarter to improve motion accuracy and response speed; in terms of mass production sprint, humanoid robots will move from the "laboratory" to the "production line", and the construction of the first production line will start simultaneously to promote the commercial implementation of products.

The commercialization path of humanoid robots is long and arduous. Only by "accumulating grain widely" through financing and "digging deep wells" with technology can practical doers wait for the dawn and witness the historic moment when humanoid robots enter factories in batches.