
The most important fact about the humanoid-robot factory UBTECH commissioned in Liuzhou is not that machines are building machines — automation has been doing that in Chinese electronics and auto plants for a decade. It is that China has now decided humanoid robots deserve the same industrial treatment: standardized lines, mixed-model assembly, and a stated production cadence, rather than hand-built prototypes rolled out for trade-show applause.
Key Points
- UBTECH’s Liuzhou plant in Guangxi has been officially commissioned as a roughly 14,000-square-meter facility designed for annual capacity above 10,000 humanoid robots.
- Reported takt time is one industrial humanoid robot completed every 10 minutes, with digital systems managing manufacturing, testing, warehousing, and shipping.
- The plant builds named, commercially positioned product lines — the Walker S series (including Walker S2) and the Cruzr series (including Cruzr S2 and Cruzr Y1) — not one-off demo units.
- Liuzhou’s factory is one of several such Chinese sites announced in the same window, including Leju Robotics/Dongfang Precision in Foshan and Xpeng’s new Iron production line, suggesting a national pattern rather than an isolated event.
- The claim rests on company and local-government statements rather than independently audited output data — standard for a commissioning announcement, and worth keeping in mind as the plant ramps toward its stated capacity.
What Was Actually Commissioned in Liuzhou
UBTECH’s Liuzhou site was formally commissioned in September, described in company and local reporting as a purpose-built industrial humanoid robot “super factory” standing 13.8 meters tall across roughly 14,000 square meters. That is not incidental detail — ceiling height and floor area determine whether a plant can run multi-story gantry systems, overhead AGV rails, and the kind of mixed-model line that lets one facility switch between different robot chassis without a full retooling. Local and national coverage frames the launch as Liuzhou’s entry into “large-scale mass production” of high-end robotics, positioning the city and Guangxi province as a manufacturing base for a category that, until recently, existed mostly in research labs and investor demo reels.
The headline production figure is a stated capacity above 10,000 units annually, with a claimed takt time — the fixed interval at which a finished unit exits the line — of one robot every 10 minutes. The plant reportedly runs a “closed intralogistics loop,” meaning parts, subassemblies, and finished units move through the facility via automated guided vehicles and digital tracking rather than manual cart-and-clipboard logistics. Reports describe robots assisting in the assembly of other robots on-site, along with digital-twin modeling and process monitoring across manufacturing, testing, warehousing, and shipping stages. The products coming off that line are not undifferentiated units: the Walker S series targets industrial and service applications, while the Cruzr line — including the S2 and Y1 variants — is built for reception, guidance, and customer-facing roles.
Why This Is Happening in China, and Why Now
Liuzhou did not emerge in isolation. China already operates roughly 2 million industrial robots in its factories — more than half the world’s total — and Shenzhen’s component supply chains let manufacturers source parts within an hour, versus over a week in much of Europe. That density of suppliers, combined with years of experience automating automotive and electronics lines, gives Chinese firms a head start in scaling humanoid hardware specifically: the joints, actuators, sensors, and battery packs a humanoid needs are largely the same components that already flow through China’s existing robotics supply base. National industrial policy has explicitly targeted humanoid robotics as a strategic sector, and Liuzhou’s local government has framed the UBTECH plant as a benchmark achievement for the region’s manufacturing upgrade.
UBTECH is not acting alone. Within the same reporting window, Leju Robotics and Dongfang Precision opened a comparable line in Foshan, Guangdong, targeting a similar 10,000-unit annual capacity at a reported 30-minute assembly interval, and Xpeng has opened its own Iron humanoid production line, with chairman He Xiaopeng noting the line was “created from scratch with no precedent to follow.” Beijing’s LY iTECH has likewise opened a dedicated humanoid super-factory in the Beijing-Tianjin-Hebei corridor. Read together, these launches describe an industry-wide shift — several major Chinese manufacturers converging, within months of each other, on the same production model: standardized humanoid lines aiming at five-figure annual output. That convergence is the more durable story here, independent of any single plant’s exact numbers.
What the Announcement Does and Does Not Establish
A commissioning ceremony marks the point at which a factory is built, equipped, and declared operational — it is the industrial equivalent of a ribbon-cutting, not a year-end production report. The figures attached to Liuzhou — 10,000-unit annual capacity, a 10-minute takt time — describe what the line is engineered to do, not a certified count of units already shipped. That distinction matters less than it might sound: capacity announcements are the normal first data point for any new industrial facility, in any country, and independent audits of steady-state throughput typically follow months or years after a plant opens, not on launch day. Comparable Chinese humanoid facilities offer a useful reference point — AgiBot’s Shanghai plant, for instance, moved from its first assembled unit in October 2024 to roughly 5,000 delivered units within about a year, before targeting tens of thousands the following year, according to production accounts from the facility itself. If Liuzhou follows a similar ramp curve, its 10,000-unit target is a milestone to track over the coming year, not a figure to expect fully realized on opening day.
It is also worth being precise about what “robots building robots” means in practice: reporting describes automated guided vehicles, robotic arms, and digital process control handling significant portions of assembly and intralogistics, alongside human oversight of testing, calibration, and final quality checks. That is consistent with how China’s existing 2-million-unit industrial robot base already operates in automotive and electronics manufacturing — heavy automation paired with human supervision at the highest-precision steps, not a fully unmanned facility.
The Broader Stakes: Manufacturing Capacity Ahead of Market Demand
The harder question facing UBTECH and its peers is not whether they can build humanoid robots at scale — Liuzhou, Foshan, and Shanghai suggest they increasingly can — but whether demand exists to absorb that output. Battery endurance still limits most humanoid units to one to four hours of active operation before recharging, and the underlying AI models still struggle to generalize across the full diversity of real factory tasks, constraints that show up consistently across independent industry analysis of China’s humanoid sector. Much of current humanoid deployment remains concentrated in reception, demonstration, and light-service roles rather than heavy industrial labor — a gap between manufacturing capacity and proven field utility that will likely define the next phase of this industry more than any single factory’s opening-day capacity claim.
For now, Liuzhou stands as concrete evidence of intent and engineering capability: a named factory, a named product line, and a stated production target, backed by a government eager to claim the milestone. Whether that capacity converts into sustained, profitable output at the scale announced is the question the next year of operations — not the commissioning ceremony — will answer.
Sources:
insiderpaper.com, pandaily.com, liuzhou.gov.cn, en.people.cn, interestingengineering.com, x.com, kantan.news, itiger.com













