Yushu Technology (宇树科技) Rushes for A‑Share "First Human Robot Stock": Behind Impressive Profits, 'Specialized' Weaknesses Need to Be Addressed

IPO push and control structure

Yushu Technology (宇树科技) is racing to become the A‑share market's first pure-play human‑robot company. The Shanghai Stock Exchange has scheduled a listing committee review for the company's initial public offering on June 1; if approved, the Hangzhou‑based hard‑tech firm plans to raise RMB 4.2 billion (42亿元) and, on the assumption of a 10% float, would sport an initial market value of at least RMB 42 billion (420亿元). Its shareholder roster—including Sequoia China (红杉), Tencent (腾讯), Alibaba (阿里巴巴), Meituan (美团) and Xiaomi‑linked Shunwei (顺为资本)—signals strong industry endorsement. Founder Wang Xingxing, however, will retain outsized control under an AB‑share voting structure, reportedly commanding about 68.78% of voting rights.

Profits built on full‑stack hardware

The numbers are eye‑catching. In 2025 Yushu recorded revenue of RMB 1.7 billion (17亿元) with non‑recurring‑adjusted net profit of RMB 590 million (5.9亿元) and gross margins above 60%. The company attributes those economics to deep vertical integration: more than 90% of core components such as motors, reducers and controllers are self‑developed, allowing hardware cost to be compressed to roughly 40% of selling price. Shipments exceeded 5,500 humanoid units in 2025, and operating cash flow was robust—RMB 670 million in 2025—suggesting the business has moved beyond pure financing dependence.

Strength in bodies, weakness in brains

But the headline profits mask a structural shortfall. Yushu's engineering emphasis has been on the robot "body" and motion control—the company's H1 humanoid has reportedly achieved running speeds above 5 m/s—while investment in large embodied AI models has lagged. R&D spend was only 8.53% of revenue in 2025 (RMB 145 million), far below peers such as UBTECH (优必选), which spent about 25%. The company has begun to remediate: it released a general VLA model, UnifoLM‑VLA‑0, in January 2026 and increased first‑quarter R&D outlays, and plans to devote roughly 85% of IPO proceeds (about RMB 3.58 billion) to model development, manufacturing scale and related projects. Can capital and time close a multi‑year technology gap?

Commercialisation, competition and geopolitical context

Yushu's commercial traction is notable—humanoids now account for over 51% of revenue and the firm has real deployments from factory floors to Tokyo's Haneda Airport—but risks remain. Early 2026 showed revenue growth alongside a sharp decline in adjusted net income, and the company itself warns that expanding scale, heavier R&D and sales spending could compress margins. Competition is intensifying: global players such as Tesla (Optimus) and Figure are accelerating hardware and AI integration, while domestic rivals are also stepping up. Amid broader US‑China tech tensions and export controls that heighten the value of domestic supply chains, Yushu's ability to convert IPO funding into competitive embodied intelligence—not just high‑volume bodies—will determine whether it evolves from a leading hardware manufacturer into a dominant intelligent‑robot ecosystem player.