NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

Wang Xingxing and Unitree: the engineer who made walking robots cheap enough to ship by the thousand

A profile of Unitree founder Wang Xingxing, whose low-cost, in-house actuation turned quadruped and humanoid robots from lab curiosities into mass-produced products.

2026-11-10 · 1019 words · NeuroAI
Wang Xingxing and Unitree: the engineer who made walking robots cheap enough to ship by the thousand

A boy who failed English so often he thought university was out of reach now runs a company that ships walking robots by the thousand.

His machines backflip on national television and pace around factory floors, priced low enough that a university lab can afford one.

The founder still signs his own engineering drawings and says the cheap version is the whole point, not a compromise.

Key takeaways

  • Wang Xingxing (王兴兴), born in 1990 in Yuyao, Ningbo, studied mechatronics at Zhejiang Sci-Tech University and Shanghai University; he did not get into Zhejiang University for graduate school after failing the English exam.
  • He built the XDog quadruped prototype on a roughly 20,000-yuan (≈ US$3,000) budget for his master's thesis, then briefly joined DJI before founding Unitree (宇树科技) in Hangzhou in 2016.
  • Unitree moved from robot dogs (Laikago, Aliengo, Go1) to humanoids — the H1 in 2023 and the G1 in 2024, the latter starting around 99,000 yuan (≈ US$13,500–16,000).
  • In 2025 humanoids became Unitree's largest revenue category, and in 2026 the company won approval to list on Shanghai's STAR Market, aiming to raise about 4.2 billion yuan (≈ US$608M).

A student who almost missed the ladder

Wang was born in 1990 in Yuyao, a county-level city in Ningbo, Zhejiang. By his own telling, English was the subject that nearly derailed him: he missed Zhejiang University and eventually earned a master's in mechatronics at Shanghai University. What he had instead was a workshop obsession. As an undergraduate at Zhejiang Sci-Tech University he built a small bipedal robot from about 200 yuan of parts, and as a graduate student he designed XDog, a quadruped driven by cheap brushless motors, on a budget of roughly 20,000 yuan (≈ US$3,000).

That thesis project became the seed of a company. Wang spent a short stint at DJI in 2016, left after about two months, and founded Unitree in Hangzhou that year. The bet was simple and unfashionable at the time: build legged robots whose core parts — motors, reducers, controllers — you make yourself, so the price can undercut everyone else.

From robot dogs to a human on two legs

Unitree's reputation was built on quadrupeds. The 2017 Laikago was its first commercial four-legged robot with self-developed core components; the 2019 Aliengo targeted industrial use, and the 2021 Go1 brought a companion robot dog to schools and hobbyists at a fraction of Western prices. The pivot to the humanoid (人形机器人) came in August 2023 with the H1, a full-size biped that walked untethered at CES 2024. In May 2024 came the G1, a compact research humanoid starting around 99,000 yuan (≈ US$13,500–16,000).

The demonstrations — running, martial arts, a yangko dance at the 2025 CCTV Spring Festival Gala performed by 16 H1 units — made the brand. But the number that matters is the one most people never see: Unitree has said humanoid sales overtook quadrupeds as its largest business in 2025, and the company reported roughly 1 billion yuan (≈ US$140M) in annual revenue by mid-2025, with its first full-year profit posted in 2024.

According to Li An, chief scientist at BrainNet (脑机网), the Unitree story is less about dancing robots and more about owning the actuation stack — motors and reducers made in-house are the real moat, not the choreography.

Why a founder stays in the workshop

Wang is the rare hardware founder who also serves as chief technology officer, and he is known for reviewing mechanical details himself. That hands-on reputation helped him personally when, in January 2025, he was seated in the front row at a high-profile business symposium hosted by China's leadership — the youngest person in that front row. Later that year Time included him in its TIME100 AI list of influential people in artificial intelligence.

Those moments sit oddly with his origin story. A founder who says he failed English exams three times in his early school years now represents a generation of Chinese hardware entrepreneurs who skipped the prestige-university path and built globally competitive products anyway.

The IPO that signals scale

In 2026 Unitree took a step most robotics startups never reach: it won approval to list on Shanghai's STAR Market. The Shanghai Stock Exchange accepted the application in March 2026 and its listing committee approved it in June, with the company aiming to raise about 4.2 billion yuan (≈ US$608M). For a firm that began as one engineer and a prototype, the listing is a marker that legged robots have moved from laboratory novelty toward a real manufacturing business.

The caveat sits in the customer mix. Many Unitree robots still go to universities, research labs, and demonstration projects rather than factory floors, and the general-purpose work — a robot that reliably does unscripted tasks — is not yet here. The motion is real; the everyday deployment is still being built.

Honest limitations

Core biographical facts (born 1990 in Yuyao, Ningbo; undergrad at Zhejiang Sci-Tech University; master's at Shanghai University in mechatronics; XDog built in grad school; founded Unitree in Hangzhou in 2016) come from profiles carried by Guangming (光明网) and encyclopedic entries. Product names, release years, and the G1 price are company-stated and widely reported. The STAR Market timeline (SSE acceptance in March 2026, committee approval in June 2026, ~4.2 billion yuan raise target) is drawn from IPO-prospectus reporting by Caixin and secondary outlets; figures are subject to change through the listing process. Revenue and profitability statements are company disclosures reported in filings and may be revised. I did not assess Unitree's safety record, overseas regulatory exposure, or the technical readiness of its general-purpose humanoid claims.

What readers can do now

  1. Separate the demo from the deployment: when evaluating any legged robot, ask for mean-time-between-human-intervention on a real task, not a choreographed clip.
  2. Watch vertical integration as the real signal — check whether a robot maker builds its own motors, reducers, and controllers in-house, because that, not the dancing, is the durable cost advantage.
  3. Follow the post-listing financials: read Unitree's revenue trend and customer mix in its disclosures rather than the first-day price move to judge whether humanoids are shipping to factories or mostly to labs.

Related coverage

Cite this article

NeuroAI (脑机网), "Wang Xingxing and Unitree: the engineer who made walking robots cheap enough to ship by the thousand," 10 November 2026, https://neuroai.site/a/na-people-wang-xingxing-unitree

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Builders” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories