November 16, 2026FundingRobotics

Weijian Technology Raises Over RMB100M Fifth Round for Construction Robots

Shanghai Weijian Technology, a construction robotics company, has closed a fifth round of financing of "over RMB 100 million", roughly US$14 million or more at current exchange rates, led by Caitong Capital, the only named investor. The news is a 36Kr exclusive posted on Nov 15, built on an interview with the company; there is no company press release, no valuation, and the amount is a range word rather than a figure, so treat the details as single-source. Report: https://36kr.com/p/4070571639067782

Weijian was founded in June 2020 and makes robots for the wet, repetitive end of building work: plastering robots (PMC-3000 and PMC-6000), spraying robots (PNP-3000 and PNP-6000), a wall-modeling robot, mortar and putty pumps, and an RMC-40 flexible rebar processing center that cuts, rolls, bends and sorts rebar with two people handling loading. Company-stated numbers: 300 to 350 square meters of plastering per robot per day with a record of 708, flatness of 2 mm over 2 m against a 4 mm spec, and a 6 m high-wall spraying robot on a Singapore project at "10x human efficiency". Founder and CEO Liang Yanxue spent about ten years at FANUC on kinematics, then worked at CATL and at Bozhilin, Country Garden's robotics arm.

The customer list is the interesting part. Weijian has a joint venture with China Construction Eighth Engineering Division, which the company says is the only robot JV approved inside the CSCEC system, and also sells to other CSCEC bureaus, Shanghai Construction Group, Japan's Takenaka, Middle East contractors including Sobha, and Singapore's HDB through centralized procurement. The company expects 2026 orders to pass RMB 100 million, more than double last year. New money goes to manufacturing, overseas expansion, and embodied-AI research.

The pitch runs against the current robotics fashion. Liang splits robot actions into "value-adding" ones like plastering, welding and coating, and "non-value-adding" ones like walking around and chatting, and says Weijian built precise motion first and only this year moved toward embodied AI, because workers on site cannot type in coordinates and process parameters. The resulting system has a "small brain" for motion control and process parameters encoded from master craftsmen, and a "big brain" for scene understanding. In a year when most robot rounds are priced on general-purpose foundation models, this is a bet that the money is in doing one dull job to spec, and that the AI layer is mainly a better user interface for the job site.
← Previous
Agents Are Pushing Code and Hardware Toward Open, and Two Essays Disagree on Why
Next β†’
Super User Daily: 2026-11-16
← Back to all articles

Comments

Loading...
>_