China's Weiyi Zhizao Secures First IPO Filing for Profitable Embodied AI Robot Maker
On Sept 18, the International Cooperation Department of the China Securities Regulatory Commission disclosed Weiyi Zhizao's overseas issuance and listing filing notice. The Changzhou-based company has cleared a precondition for a Hong Kong Stock Exchange hearing.
The name is unfamiliar to most, but its industry position deserves attention: it is China's largest industrial embodied AI robot product provider and ranks first by market share under IDC's measure. Embodied AI is the hottest sector this year, with Unitree, Zhiyuan, and Galbot queuing at billion-yuan valuations. Yet the first ticket to a profitable post-IPO has gone to a robot company that does not dance at launch events.
Dancers queue; screwdrivers land first. On the humanoid side, Unitree Technology's STAR Market IPO took only 73 days from acceptance to approval. Zhiyuan Robot headed to HKEX after share reform, while Galbot and Galaxea were valued above 20 billion yuan. In the first half, the sector saw 288 financings raising more than 46 billion yuan.
Already-listed robotics companies UBTech, Dobot, Geek+, and Yunji share one trait: they are not yet profitable. More than 30 robotics supply-chain companies are queuing on HKEX, and most prospectuses carry the same phrase: strategic loss. Weiyi is the counterexample.
Its financials are not glamorous. Revenue in 2025 was 796 million yuan, and net profit was only 5.06 million yuan—just above the profit line. But it holds three things peers lack: two consecutive years of profitability, positive operating cash flow of 141 million yuan, and a No. 1 position validated by factories rather than stories.
While humanoid robots compete on dance moves, Weiyi's robots have spent eight years doing inspection, polishing, assembly, and loading and unloading on production lines.
The passing line is thousand-unit real deployment. In 2018, Pan Zhengyi founded the company in Changzhou, starting with AI visual inspection. In 2022, it acquired Zhiyuntiangong, founded by Zhang Zhiqi, a 1970s-born executive with 16 years at SAP, former vice president of Baidu Intelligent Cloud, and former vice president of DingTalk, who then took the helm.
Zhang said in a recent interview that the passing line for industrial embodied AI is thousand-unit real deployment—not laboratory demonstrations, but thousands of machines working every day on real production lines. The moat lies in data. The prospectus discloses more than 23TB of real industrial precision-labeled data, while CIC statistics show most companies had less than 5TB in 2024.
Its core EIIR revenue grew from 114 million yuan in 2023 to 453 million yuan in 2025, rising from 26.3% to 57% of total revenue, with a gross margin of 53.5%.
There are three hurdles. The first is the thinness of profit: net profit fell from 15.74 million yuan in 2024 to 5.06 million yuan in 2025, just above the line, while R&D spending is 200–300 million yuan a year, nearly 30% of revenue, and computing costs doubled in a year.
The second is customer and supplier concentration: the top five customers account for 38.1% of revenue, the top five suppliers 61.7% of procurement, and the largest supplier 26.9%.
The third is governance. After the 2022 acquisition, Zhang held about 14.23% and founder Pan held 17.55% through a holding platform. The larger shareholder ceded decision-making power to the smaller one, with a concert-party agreement stating Zhang prevails in disagreements. This inverted equity and control structure, plus Pan's 2024 capital-withdrawal dispute and early investor Baidu's pre-IPO exit, will be unavoidable issues in the hearing.
The real weight of this ticket is that it may change the pricing logic of embodied AI. Over the past year, the sector's valuation anchor has been narrative: Unitree's listing, Zhiyuan's acquisition, and more than 20 members of the billion-yuan club, all valued on imagined space. Weiyi offers a new reference point: what is a company with 800 million yuan in revenue, 5 million yuan in net profit, the No. 1 market share, and positive cash flow worth?
Once it lists, the market's ruler may shift from how high it can jump to how much it can earn. It also answers where embodied AI commercialization will first explode. Humanoid advocates point to general scenarios; Weiyi's financials show the factory. Inspection, polishing, and loading and unloading—dull but essential—are where AI can first calculate a return on investment.
Three points are worth watching: whether the hearing is passed, what valuation the IPO pricing gives, and the ratio of institutions to retail investors in the subscription.