Chinese Embodied AI Firm Secretly Files for HK IPO, Readies Q4 Robot Launch
Zidongliang Robot has secretly filed for an IPO on the Hong Kong Stock Exchange, according to sources familiar with the matter. The company is the only player in the embodied intelligence field that has built a fully self-developed end-to-end embodied model from scratch, while many competitors merely fine-tune open-source models and repackage them as proprietary.
The startup's core differentiator is its "World Unified Model" (WUM) architecture, which integrates native multimodal capabilities from the start, eliminating the information loss inherent in conventional VLA models that chain separate modules. In May, it open-sourced Wall-OSS-0.5, the world's first embodied foundation model that can be deployed directly to real robots after pretraining without task-specific fine-tuning. This marks a significant leap from modular VLA pipelines to a unified native-multimodal framework.
Behind this innovation is a robust engineering system. Founder Wang Qian established China's first large-scale robot data factory in Shenzhen. In July, the company launched the QUANXTA Zero series, a data collection solution designed inversely from the actual needs of model training, covering the entire pipeline including collection, cleaning, auto-labeling, and quality control.
These efforts have attracted massive capital. Between late April and June 2026, the company closed four consecutive funding rounds (B, B+, B++, and C), with its valuation surging from RMB 10 billion to over RMB 20 billion. The rounds were led by four internet giants—Meituan, Alibaba, ByteDance, and Xiaomi—with participation from top-tier VCs such as Sequoia China and IDG Capital, as well as state-owned investors including China Mobile and China Development Bank Finance.
According to investment bankers, the company's valuation of RMB 20 billion could make future fundraising more challenging, making the IPO a natural step to open wider financing channels. Meanwhile, the third-generation robot is scheduled for release in Q4, as the company accelerates its dual-track strategy of technological innovation and capital expansion.