In the embodied AI track of 2026, the strategy of 'fundraising while investing' has evolved from a special tactic for a few leading companies to a norm for participants in the sector. IT桔子 has previously conducted a special report and analysis on this phenomenon. Data shows that at least 29 domestic embodied AI companies are currently making external investments.
Today, we focus on Leju Robot, which has relatively high activity in this regard.
I. About Leju:
A Humanoid Robot OEM Charging Towards an IPO
In 2016, Leng Xiaokun founded Leju in Shenzhen with partners from the HIT Robot Club. A decade later, they developed the humanoid robot "Kuafu"—China's first full-size, highly dynamic humanoid robot capable of jumping, adapting to multi-terrain walking, and among the first to be equipped with OpenHarmony and integrated with Huawei's Pangu model.
According to data from institutions like Omdia and IDC, Leju's shipment volume ranked among the top four globally in 2025, with its full-size bipedal robot ranking second globally. In the same year, 577 Kuafu units were sold, generating revenue of approximately 178 million RMB (a roughly 12-fold year-on-year increase), accounting for nearly 70% of the company's revenue. Overall revenue grew from 54 million RMB in 2023 to 258 million RMB in 2025.
Behind the impressive shipment growth lies profitability pressure: the company accumulated net losses exceeding 170 million RMB over three years, with gross margin sliding from 50.45% to 40.78%, exhibiting characteristics of 'trading price for volume'.
In May 2026, Leju's ChiNext IPO application was accepted by the Shenzhen Stock Exchange (the first to apply under the 'fourth set of standards'), aiming to raise 2.6 billion RMB with a post-investment valuation of 4.327 billion RMB.
The prospectus reveals its industrial logic: on one hand, advancing core component localization and supply chain collaboration for cost reduction; on the other, listing the embodied AI R&D center and high-quality dataset as key fundraising projects—this set of 'cost reduction, full-stack capability, and capacity expansion' tactics is almost precisely mirrored in the targets of its external investments over the past two years.
II. Panorama of 13 External Investments:
Three-Tier Positioning Along the Value Chain
According to IT桔子 data, as of August 2026, Leju has made 13 investments as an investor, which can be categorized into three types based on the value chain: core component supply (4), software & intelligence layer (6), and scenario joint ventures & industrial synergy (3).
III. Logic Analysis:
How Were the 13 Investments Made?
1. Core Components: Keeping Cost Control in Its Own Hands
The most expensive parts of a humanoid robot are precisely the motion joints and end effectors, and the prospectus precisely lists 'supply chain collaboration for cost reduction' as the fundamental source of Kuafu's price advantage.
Leju's four investments in components equate to grasping the lifeline of cost reduction within its shareholder registry.
The most notable is Lingxinqiaoshou. This Beijing Haidian-based company is already a unicorn in the dexterous hand sector, securing hundreds of millions in its A+ round.
Dexterous hands are widely recognized as the end effector, the 'final piece of the puzzle' for humanoid robots, directly determining whether a robot can perform delicate tasks like grasping, pinching, and twisting—Leju's bet on the leader secures both the supply and pricing power of high-end dexterous hands, while ensuring Kuafu does not rely on external purchases for 'real work', aligning highly with the industrial and commercial service applications emphasized in its prospectus.
Further upstream, Quanzhibo and Lijudongli are positioned in joint modules and joint motors, respectively.
Quanzhibo, based in Wuxi, focuses on robot joints and servo motors, and is already a high-tech enterprise; Lijudongli, located in Ningbo, specializes in joint motors.
Together, these two form a 'module + drive' combination, allowing Leju to control the heaviest part of its cost structure and avoid the earlier predicament of 'relying on imported core components and spending millions to build a prototype generation'.
Heiman Technology adds a future option for bionic core components, leaving room for differentiation in the next-generation platform.
2. Software & Intelligence Layer: Creating Backups for the 'Brain'
Leju's strength lies in the 'cerebellum'—high-dynamic bipedal motion control, while the 'brain' has long relied on external partners like Huawei's Pangu. The six investments in the software layer represent multiple backup forces it has positioned on the 'brain' side.
The youngest and most aggressive of these is Deta Intelligent.
Founded only in 2026, with a team boasting Google and Facebook backgrounds, this company raised nearly 100 million RMB in its seed round. Its bet is on the strategic high ground of 'humanoid foundational models', backed by the imaginative space of world models and spatial intelligence.
Correspondingly, Jubao Panshi focuses on developing the 'embodied brain' for humanoid robots, with a team carrying Huawei backgrounds—while deeply integrating with HarmonyOS and Pangu, Leju clearly intends to maintain a self-controllable backup, leveraging the ecosystem's light while not putting all its eggs in one basket.
data-check-id="558120">Jushen Data and Yunkexing extend the reach into training data.The former works on monetizing data assets for embodied AI, while the latter is a sector-specific scenario data platform with a team carrying Apple backgrounds.
Both directly target the 'high-quality, large-scale dataset' (approximately 616 million RMB) listed in this IPO's fundraising projects, effectively extending the R&D system externally to supply data for Kuafu's continuous iteration.
Additionally, Xingyuan Zhi Robot, Xuanji Intelligent, etc., also have their stakes in general embodied brains and multimodal vision, making this 'external brain' network denser.
3. Scenario & Industrial Synergy: From Selling Devices to Selling Solutions
The remaining three investments are not purely in supply, pointing towards real-world deployment—what Leju seeks is not just financial returns, but channels to transform Kuafu from 'a single robot' into 'a set of solutions'.
Land-Sea New Corridor is a cross-border logistics service provider with nearly a decade of experience. Leju's investment opens real channels for humanoid robots to enter logistics handling, and even follow Chinese manufacturing overseas, echoing the 'logistics' industrial scenario specified in the prospectus.
Jushi Intelligent and Wushi Innovation complement the whole-machine ecosystem: the former is a Shanghai-based company also developing humanoid robots, serving as an ecosystem synergy and co-opetition partner for complementary whole-machine solutions and joint delivery; the latter, rooted in Shenzhen developing service robots, neatly fits Leju's scenarios like exhibition hall guiding and storefront shopping assistance.
By pulling upstream and downstream scenario partners into its circle through investments, Leju's industrialization narrative becomes complete.
IV. The Ecosystem Empire and Underlying Concerns
As humanoid robotics approaches the eve of industrialization, simply 'self-developing the platform' is no longer sufficient to constitute a barrier.
Leju's choice is to use capital to 'weave' the value chain into a network—self-developing the body, externally connecting the brain, holding stakes in the supply chain. This not only tells a complete industry chain story for the IPO but also lays the groundwork for mass production cost reduction and scenario deployment.
However, the concerns are equally clear: Leju itself remains loss-making, with operating cash flow persistently negative and gross margins declining; most of the 13 investments are early-stage, with undisclosed or small amounts, long return cycles, and most have not yet formed quantifiable synergistic revenue with the parent entity.
If the industry experiences short-term overcapacity and intensified price wars, the 'synergistic dividends' of this ecosystem network may not materialize as expected.
Founder Leng Xiaokun once stated frankly: 'Only doing the dirty and hard work constitutes a barrier; technology may not necessarily become the industry's barrier.' Leju is using 13 external investments to gradually turn those 'dirty and hard jobs'—joints, motors, dexterous hands, embodied data—into its own moat.
Whether this network can ultimately catch the future of a quasi-listed company, time will tell.
This article is from the WeChat public account: IT桔子 , Author: Judy
