广告
加载中

前字节专家加盟 这家绳驱机器人再获10亿融资

亿邦动力 2026-10-14 11:15
亿邦动力 2026/10/14 11:15

邦小白快读

EN
全文速览

这篇文章介绍了星尘智能这家机器人公司,普通人可以关注两件事:可以去哪里体验机器人,以及未来家用机器人会是什么样。

1.体验渠道:公司自主研发的机器人零售服务店“机器人MART”已在北京、上海、广州等六个城市批量化部署,机器人从单点展示变成持续运营,普通人可以在这些门店直观看到机器人干活。

2.现有能力:机器人应用覆盖零售、文旅、工业、科研、教育等场景,客户包括央视、京东等;新一代人形机器人T1已于2026年6月开始发货,将高精细操作机器人带到9万元价格区间,说明产品已走出实验室。

3.未来设想:面向家庭和个人用户的系列产品正在研发,同时公司计划拓展具身操作系统AOS的软件及应用服务,意味着未来机器人可能像智能手机一样承载各种应用。

星尘智能的产品定价、零售渠道和家庭化布局,给品牌商在机器人新消费品类上提供了参考。

1.渠道建设:机器人零售服务店“机器人MART”在六个城市批量部署,让机器人从展示走向持续运营,这是一种品牌自建体验与零售一体化的新渠道形态。

2.定价与价格竞争:新一代人形机器人T1将高精细操作机器人带到9万元价格区间,同时拉通万台级产能,用价格下探和成本控制抢占市场,为同类品牌设定竞争坐标。

3.产品研发方向:以AI模型、具身OS和绳驱本体构筑全栈能力,通过零售、导览、艺术表演等场景方案积累真实运营数据和用户反馈,为向家庭服务场景拓展打基础。

4.消费趋势观察:机器人正从商业场景走向家庭;面向家庭和个人用户的产品已在研发,未来软件应用服务可能成为新的商业模式。

星尘智能的进展为卖家提供了商用机器人市场的增长信号和合作线索。

1.增长市场:应用覆盖零售、文旅、工业、科研、教育等场景,与金马游乐、仙工智能合作将部署上千台,与中科创达推动千台级落地,B端采购需求明确。

2.合作机会:公司沿商业服务和开发者生态双路径推进,提供标准化机器人平台、技能模块与开发工具;同时借助中科创达覆盖北美、欧洲、亚洲的全球销售网络出海,卖家可关注渠道合作。

3.风险提示:绳驱机器人量产存在绳索弹性蠕变、张力控制难、寿命磨损、批量一致性等难题,合作时需评估产品可靠性和技术积累。

4.学习点:软硬件强耦合的全栈自研、从原型到规模量产、9万元价格带,是新品类打法的参考。

绳驱机器人的量产进程为工厂带来零部件与工艺的新需求,其工程方法也有借鉴价值。

1.产品设计需求:绳驱本体用绳索、卷筒和滑轮替代部分刚性连杆和齿轮,适合灵巧手、仿生机器人和外骨骼;腱绳材料、张紧控制、卷筒滑轮等部件存在配套机会。

2.生产难点与解决:绳索弹性蠕变和形变会导致定位漂移和刚度不足,张力控制难,预紧不当会加剧摩擦磨损,同时面临装配标定、批量一致性、长期维保等问题。星尘智能用自研补偿算法配合材料与张紧控制,并把材料、结构、控制、标定、验证和维保串起来,值得工厂参考。

3.商业机会:已拉通万台级产能,新一代人形机器人T1开始发货,9万元价格区间带动量产需求,上游零部件制造有空间。

4.数字化启示:软硬件强耦合和全栈自研模式说明数字化能力需与机械设计紧密协同,部署中积累数据持续改进的逻辑可用于工厂数字化建设。

星尘智能展现了具身智能服务领域的技术趋势与服务机会。

1.行业趋势:具身智能走向全栈自研与软硬件强耦合,模型、操作系统与本体协同成为主流方向;融资频次和估值增长表明赛道升温。

2.新技术与客户痛点:绳驱驱动灵活精细,但量产存在绳索弹性蠕变、形变、定位漂移、张力控制难、寿命短、批量一致性差等问题。

3.解决方案:通过自研补偿算法建模与补偿弹性和形变,在腱绳材料、张紧控制上协同设计,并完成装配标定、可靠性验证与批量一致性工程积累,可作为服务商为客户提供方案的参考。

4.服务机会:商业服务和开发者生态双路径落地,零售、文旅、工业、科研、教育等场景都需要部署、运维、应用开发服务;机器人MART门店持续运营也带来新服务需求。

星尘智能在零售平台和软件平台上的探索,为平台商提供了运营与生态建设参考。

1.商业模式:商业服务与开发者生态双路径并行,商业服务把本体、模型和AOS整合为零售、导览、艺术表演等方案;开发者生态为开发者和机构提供标准化平台、技能模块与工具。

2.平台做法:机器人MART作为机器人零售服务店,已在六个城市批量化部署,从单点展示走向持续运营,是自营体验与机器人服务结合的新形态。

3.运营管理:通过真实运营验证产品能力、积累用户反馈,并沉淀可复用交互、操作与安全控制能力,为向家庭场景扩展打底;未来计划拓展AOS软件及应用服务,探索模型商业化。

4.风险规避:绳驱机器人量产有可靠性和一致性风险,平台在推广机器人应用时需关注设备维护与服务网络;需借助生态伙伴的行业知识拓展应用,也要谨慎选择合作方。

星尘智能案例是研究具身智能资本、技术、商业模式和人才流动的新样本。

1.产业新动向:半年内完成4轮融资,估值达到150亿元,股东结构涵盖国资基金、产业集团与头部财务机构;前字节AI专家孙鹏加盟,显示人才从互联网大厂流向机器人企业。

2.技术路线与研究问题:绳驱本体轻量灵活,但量产存在绳索弹性蠕变、张力控制、磨损寿命、装配标定与批量一致性等难题;自研补偿算法与工程积累是可行的解决路径。

3.商业模式:商业服务和开发者生态双路径,机器人MART零售店形成部署与运营闭环;与多家伙伴签署千台级订单,依托全球销售网络出海;家庭场景和AOS软件服务商业化仍在探索。

4.启示:全栈自研形成本体+数据+模型的数据闭环,从原型验证到规模量产的能力积累,对其他企业有借鉴意义;量产可靠性和家庭场景安全控制需求,为相关法规标准建设提供现实参照。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article introduces the robotics company Astribot. For general readers, two things are worth watching: where to experience the robots, and what future home robots will look like.

1. Experience channels: The company's self-developed robot retail service stores, "Robot MART," have been deployed at scale in six cities including Beijing, Shanghai, and Guangzhou. Robots have moved from single-point displays to continuous operations, and ordinary people can watch robots working directly in these stores.

2. Current capabilities: Robot applications cover retail, cultural tourism, industry, scientific research, education, and other scenarios, with customers including CCTV and JD.com. The new generation humanoid robot T1 began shipping in June 2026, bringing high-precision manipulation robots to the RMB 90,000 price range and showing that the product has moved beyond the lab.

3. Future vision: Product series for households and individual users are under development, and the company plans to expand software and application services for its embodied operating system AOS. This suggests robots may eventually carry diverse applications just like smartphones.

Astribot's product pricing, retail channels, and home-oriented layout provide brand owners with a reference for positioning robots as a new consumer category.

1. Channel development: The robot retail service store "Robot MART" has been deployed at scale in six cities, shifting robots from display to continuous operation. This is a new self-built channel format that integrates experience and retail.

2. Pricing and price competition: The new generation humanoid robot T1 brings high-precision manipulation robots into the RMB 90,000 price range while also reaching production capacity at the 10,000-unit level. By lowering prices and controlling costs, it is capturing market share and setting the competitive benchmark for similar brands.

3. Product R&D direction: The company builds full-stack capabilities through AI models, the embodied OS, and cable-driven hardware. It accumulates real operational data and user feedback through solution scenarios such as retail, guided tours, and artistic performance, laying the groundwork for expansion into home service scenarios.

4. Consumer trend observation: Robots are moving from commercial settings to households. Products aimed at households and individual users are already in R&D, and software application services may become a new business model in the future.

Astribot's progress offers sellers growth signals and cooperation leads in the commercial robot market.

1. Growing market: Applications cover retail, cultural tourism, industry, scientific research, education, and other scenarios. Through partnerships with Jinma Amusement and SEER, more than a thousand units will be deployed; with ThunderSoft, the company is pushing toward thousand-unit-scale adoption. Demand from business buyers is clear.

2. Cooperation opportunities: The company is advancing along dual paths of commercial services and a developer ecosystem, offering standardized robot platforms, skill modules, and development tools. It is also going global through ThunderSoft's sales network covering North America, Europe, and Asia. Sellers should watch for channel partnership opportunities.

3. Risk warning: Mass production of cable-driven robots still faces challenges such as elastic creep in cables, difficult tension control, wear and lifespan issues, and batch consistency. When cooperating, sellers should evaluate product reliability and the company's technical track record.

4. Takeaways: Full-stack in-house development with tight software-hardware coupling, the transition from prototype to mass production, and the RMB 90,000 price point are useful references for a new-category playbook.

The mass production of cable-driven robots creates new demand for parts and processes among factories, and its engineering methods are also worth referencing.

1. Product design requirements: The cable-driven body replaces some rigid links and gears with cables, drums, and pulleys, making it suitable for dexterous hands, biomimetic robots, and exoskeletons. There are supporting opportunities in tendon cable materials, tension control, drums, and pulleys.

2. Production difficulties and solutions: Elastic creep and deformation of cables can cause positioning drift and insufficient stiffness; tension control is difficult, and improper pre-tensioning can worsen friction and wear. The process also faces assembly calibration, batch consistency, and long-term maintenance issues. Astribot addresses these with self-developed compensation algorithms combined with material and tension control, and integrates materials, structure, control, calibration, verification, and maintenance into one chain. This is a useful reference for factories.

3. Business opportunities: The company has reached 10,000-unit-scale production capacity. The next-generation humanoid robot T1 has begun shipping, and the RMB 90,000 price point is driving mass-production demand, leaving room for upstream parts manufacturing.

4. Digitalization insight: The tightly coupled software-hardware, full-stack in-house model shows that digital capabilities must closely coordinate with mechanical design. The logic of accumulating data during deployment and continuously improving can be applied to factory digitalization.

Astribot illustrates the technology trends and service opportunities in the embodied intelligence services field.

1. Industry trend: Embodied intelligence is moving toward full-stack in-house development and tight software-hardware coupling. The convergence of models, operating systems, and hardware bodies is becoming the mainstream direction; increasing funding frequency and valuation growth show the sector is heating up.

2. New technologies and customer pain points: Cable-driven actuation is flexible and precise, but mass production faces issues such as cable elastic creep, deformation, positioning drift, difficult tension control, short lifespan, and poor batch consistency.

3. Solutions: The company uses self-developed compensation algorithms to model and compensate for elasticity and deformation, coordinates tendon cable materials with tension control, and builds engineering capabilities in assembly calibration, reliability verification, and batch consistency. This can serve as a reference for service providers designing solutions for clients.

4. Service opportunities: The dual-path focus on commercial services and a developer ecosystem is being implemented on the ground. Retail, cultural tourism, industry, scientific research, education, and other scenarios all require deployment, maintenance, and application development services. The continuous operation of Robot MART stores also generates new service demand.

Astribot's exploration of retail platforms and software platforms offers marketplace operators a reference for operations and ecosystem building.

1. Business model: It advances in parallel with a commercial services track and a developer ecosystem track. Commercial services integrate hardware, models, and AOS into solutions for retail, guided tours, artistic performance, and more; the developer ecosystem provides developers and institutions with standardized platforms, skill modules, and tools.

2. Platform approach: Robot MART, as a robot retail service store, has been deployed in six cities at scale and has moved from single-point displays to continuous operations. It is a new form that combines self-operated experience with robot services.

3. Operations management: Through real operations, the company validates product capabilities, accumulates user feedback, and builds reusable interaction, operation, and safety-control capabilities, laying a foundation for expansion into home scenarios. It also plans to expand AOS software and application services and explore model commercialization.

4. Risk mitigation: Cable-driven robot mass production carries reliability and consistency risks. When promoting robot applications, platforms need to pay attention to equipment maintenance and service networks. They also need to leverage ecosystem partners' industry knowledge to expand applications and choose partners carefully.

The Astribot case is a new sample for studying capital, technology, business models, and talent flows in embodied intelligence.

1. New industry dynamics: The company completed four financing rounds in six months and reached a valuation of RMB 15 billion. Its shareholder base includes state-backed funds, industrial groups, and leading financial institutions. Former ByteDance AI expert Sun Peng has joined, showing talent moving from internet giants to robotics companies.

2. Technology route and research questions: The cable-driven body is lightweight and flexible, but mass production faces challenges such as cable elastic creep, tension control, wear and lifespan, assembly calibration, and batch consistency. Self-developed compensation algorithms combined with engineering accumulation are a viable solution path.

3. Business model: The company operates a dual path of commercial services and a developer ecosystem. Robot MART retail stores create a closed loop of deployment and operations. It has signed thousand-unit-scale orders with multiple partners and is going global through a worldwide sales network. The commercialization of home scenarios and AOS software services is still being explored.

4. Implications: Full-stack in-house development creates a data closed loop spanning hardware, data, and models, and the ability accumulated from prototype verification to mass production is instructive for other companies. The need for reliable mass production and safety control in home scenarios provides a practical reference for relevant regulations and standards.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

【亿邦原创】日前,星尘智能完成10亿元规模新一轮融资,估值达150亿元。本轮融资由投控东海(深投控)领投,洪泰基金、光华阳光管理的宁海兴泰合基金、善恒资本等国资及财务投资人参与投资,老股东先沐追投。

这已是星尘智能近半年内完成的第4轮融资。

自2026年3月以来,星尘智能已先后完成A+轮、A++轮、B轮及本轮融资。其中,星尘智能在2026年6月完成B轮系列融资总额超10亿元、估值首次突破百亿元;仅3个月后便宣布完成新一轮融资、估值达150亿元。此前,星尘智能已先后获得云启资本、经纬创投、字节系锦秋基金、蚂蚁集团,国科投资等参与投资。本轮融资后,星尘智能股东结构已构建"国资基金+产业集团+头部财务机构"的多元优质股东结构。

星尘智能成立于2022年12月,基于“Design for AI”核心理念,构建了“AI模型—具身OS—绳驱本体”三位一体的全栈自研体系。围绕Lumo具身基座模型、Astribot OS与绳驱本体全栈自研,构建了模型、操作系统与本体协同的三位一体全栈架构,形成了“本体+数据+模型”的技术闭环。

如何破解绳驱量产难题?

提到星尘智能,外界想到的第一个标签就是“绳驱”。

绳驱作为机器人驱动方式的一种,是用绳索、卷筒和滑轮替代部分刚性连杆和齿轮,驱动器收放绳索,经滑轮导向拉动关节。由于绳索只能拉、不能推,通常需要双绳对拉,或借助弹簧、重力回位。它的优点是轻量、布局灵活,可以把电机放在远端、降低末端惯量,适合灵巧手、仿生机器人和外骨骼等场景。

星尘智能的机器人本体就是采用了此种驱动方式。在星尘智能创始人、CEO来杰看来,绳驱技术为机器人提供灵活、精细且更适合人机交互的执行能力、为模型训练提供高质量数据;Lumo模型负责理解环境、推理和生成动作;Astribot OS把这些能力组织成用户和开发者能够调用的应用。三者协同,使机器人既能完成实际任务,也能利用部署中积累的数据持续改进。

但绳驱量产并不容易。

绳索有弹性,还会蠕变和形变,绳长会随载荷和时间变化,导致定位漂移和刚度不足。张力控制也很难,预紧不足会空程,预紧过大又会加剧摩擦和磨损。反复弯曲、摩擦和冲击还会缩短寿命。装配标定、批量一致性和长期维保同样关键,实验室能实现,不代表批量产品能稳定复现。

为了解决这些难题,星尘智能选择软硬件强耦合的方式。它通过自研补偿算法,对绳驱的弹性、蠕变和形变进行建模与补偿,并在腱绳材料、张紧控制上配合设计,减少松弛、摩擦和磨损。同时,它在装配标定、可靠性验证和批量一致性上形成工程积累,把材料、结构、控制、标定、验证和维保串起来。

凭借绳驱技术的全栈积累,星尘智能完整跑通了从原型验证到规模量产的发展路径,展现出极强的产品迭代与成本控制的稀缺性。2025年9月,星尘智能S旗舰系列开启千台级订单,2026年6月其新一代人形机器人T1开始发货、率先将高精细操作机器人带到9万元价格区间实现,叠加其同步拉通的万台级产能,使其在技术代差、产品普及速度与成本壁垒上进一步拉开代际差距。

前字节专家加入星尘

在团队方面。星尘智能创始人兼CEO来杰,在人工智能与机器人领域拥有超过17年研发设计经验,主导研发过多款新型机器人,此前他曾担任腾讯机器人实验室1号员工与架构师、百度“小度机器人”团队负责人,持续推动AI机器人助理从梦想变为现实。同时,星尘智能核心创始团队共6人均出自腾讯Robotics X实验室,成员背景覆盖腾讯、谷歌、优必选、百度、华为等前沿科技企业。

公开信息显示,国内领先AI人才孙鹏于今年9月正式加盟星尘智能,将重点负责机器人强化学习方向的技术研发与应用探索。

在加入星尘智能之前,他先后在腾讯AI Lab和Robotics X任职,从事深度强化学习与机器人控制研究,并主导研发《星际争霸》AI智能体TStarbots、TStarbotX;后在字节跳动游戏AI部门及字节AI Lab/ByteResearch任职,负责多智能体与分布式大规模强化学习,主导开发ByteRL,参与发布ReFT及DeltaProver,并深度参与大模型预训练、RLHF与模型对齐。

即将迈向家庭场景

在商业场景落地方面,星尘智能应用覆盖零售、文旅、工业、科研、教育等场景。客户包括金马游乐、中科创达、博登智能、仙工智能、央视及京东等,包括与金马游乐、仙工智能等合作将在商业与工业等场景部署上千台AI机器人,并将与中科创达共同推动千台级规模落地,依托后者覆盖北美、欧洲、亚洲的全球销售网络,加速推进机器人产品出海,进行全球化布局。

由星尘智能自主研发的机器人零售服务店“机器人MART”自2025年11月起陆续在北京、上海、广州等六个城市批量化部署,推动机器人从单点展示走向持续运营。

在商业化方面,星尘智表示,未来能将继续沿“商业服务”与“开发者生态”双路径推进商业化,面向商业服务市场,星尘智能将机器人本体、AI模型与具身操作系统AOS的全栈能力整合为零售、导览、艺术表演等场景解决方案,通过真实运营验证产品能力、积累用户反馈,并沉淀可复用的交互、操作与安全控制能力,为向家庭服务场景拓展打下基础。面向科研机构、AI及行业应用开发者,星尘智能提供标准化机器人平台、技能模块与开发工具,降低机器人应用开发门槛,借助生态伙伴的行业知识拓展产品能力与应用场景。

作为长期产品布局的重要方向,星尘智能面向家庭和个人用户的系列产品已正在研发阶段。而随着装机量和使用频率增长,星尘智能计划拓展具身操作系统AOS的软件及应用服务,持续探索模型的商业化模式。

本文首发于 亿邦动力 官方网站

文章来源:亿邦动力

广告
微信
朋友圈

FAQ回顾

什么是绳驱机器人?它和传统机器人有什么不同?

绳驱机器人是采用绳驱驱动方式的机器人,用绳索、卷筒和滑轮替代部分刚性连杆和齿轮,驱动器收放绳索经滑轮导向拉动关节。其优点是轻量、布局灵活,可将电机放在远端降低末端惯量,适合灵巧手、仿生机器人和外骨骼等场景。星尘智能是国内代表性绳驱机器人企业,其机器人本体即采用这种驱动方式。

星尘智能是做什么的?

星尘智能成立于2022年12月,是一家基于“Design for AI”理念的具身智能机器人公司,构建了“AI模型—具身OS—绳驱本体”三位一体全栈自研体系,围绕Lumo具身基座模型、Astribot OS与绳驱本体全栈自研,形成“本体+数据+模型”技术闭环,产品覆盖零售、文旅、工业、科研、教育等场景。

绳驱机器人量产难点有哪些?如何解决?

绳驱量产难点包括:绳索弹性、蠕变和形变导致定位漂移和刚度不足;张力控制难,预紧不足会空程、过大加剧摩擦磨损;反复弯曲摩擦冲击缩短寿命;装配标定、批量一致性和长期维保难以稳定复现。星尘智能通过自研补偿算法对弹性和蠕变建模补偿,在腱绳材料、张紧控制上配合设计,并将材料、结构、控制、标定、验证和维保串联,跑通了从原型验证到规模量产路径。

星尘智能近期融资情况如何?

2026年10月,星尘智能完成10亿元新一轮融资,估值达150亿元,由投控东海领投,洪泰基金、宁海兴泰合基金、善恒资本等参与,老股东先沐追投。自2026年3月以来,星尘智能已完成A轮、A+轮、B轮及本轮共4轮融资;其中2026年6月B轮系列融资总额超10亿元,估值首次突破百亿元。此前投资方还包括云启资本、经纬创投、字节系锦秋基金、蚂蚁集团、国科投资等。

星尘智能的机器人产品在商业化方面有哪些进展?

星尘智能应用已覆盖零售、文旅、工业、科研、教育等场景,客户包括金马游乐、中科创达、博登智能、仙工智能、央视及京东等。其S旗舰系列于2025年9月开启千台级订单,2026年6月新一代人形机器人T1开始发货,将高精细操作机器人带到9万元价格区间,并同步拉通万台级产能。此外,公司计划与金马游乐、仙工智能等部署上千台AI机器人,并与中科创达推动千台级规模落地及海外市场拓展。

这么好看,分享一下?

朋友圈 分享

APP内打开

赞 +1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0