Startup 纽娲机器人 bets on World Traversal Model for last-mile delivery
A 钛媒体 authorized interview with founder 杨睿刚 profiles 纽娲机器人, founded in early 2026, which targets the last few hundred meters of unmanned delivery by having robots enter residential compounds, buildings and elevators to reach apartment doors. The report says its in-house “World Traversal Model,” trained on data from a proprietary simulation engine, can run on multiple delivery robot types, and that deployment requires no prior mapping or building modification, with a compound usually deployed in one day and running within two; those capability claims come from the company and were not independently verified. It also reports a 50 million yuan angel round in June with 蓝湖资本 named as lead investor, an August product launch at the World Robot Conference, a self-operated human-robot hybrid station in 桐庐 opened August 1, a UAE debut at Abu Dhabi’s LIVEX exhibition in late September, and a PoC with Emirati state-owned AI company K2 that the report says landed in a high-rise apartment building at Abu Dhabi’s Department of Municipalities and Transport (DMT). It cites plans for three more PoC sites by year-end, hundreds of deployments next, and scale deployment from 2028.
The report argues that the last few hundred meters into spaces managed by property firms and owners is the part of unmanned delivery that has resisted scale for a decade, with deployment mapping and building retrofits a core cost, so 纽娲’s map-free World Traversal Model and its 桐庐 and Abu Dhabi trials are an early market test of that approach.
(本文作者为 科技异端,钛媒体经授权发布)

物流快递的无人化进程已走了十年,从京东推出国内第一辆无人配送车算起,美团、菜鸟、百度、顺丰先后进场,一批创业公司陆续入局。到今天,根据公开资料,全国物流无人车保有量超过3.5万台,200多个城市开放路权,中通的无人车队每天运送超过20万件包裹。
而随着开放无人配送的路段越来越多,靠近家门的那几百米,却成了整个行业最难攻克的堡垒。进小区、进单元楼、上电梯、把东西放到用户门口,这些事至今还是快递员在做,更多的时候,是消费者自己去驿站和快递柜取。
过去十年,尝试突破这最后几百米的公司很多,但大多数都没了下文:京东的配送机器人曾一度送进过30多个城市的校园、园区和街道,如今却依然选择退守驿站;菜鸟的小蛮驴实际配送跑了多年,今年年初,整个无人车业务并入九识智能;顺丰等头部企业也在楼宇场景探索多年,入户配送至今仍未形成规模。
#纽娲机器人#杨睿刚#K2 Group#蓝湖资本#Plug and Play China#World Traversal Model#AutoGo#LIVEX#World Robot Conference#桐庐人机混合驿站