• 亿邦公众号
    • APP下载

融了10个亿 又一百亿独角兽诞生

作者:黎曼 文章来源:投中网

2026-07-20 04:40

作者:黎曼 文章来源:投中网

2026-07-20 04:40

邦小白快读

EN
全文速览

本文核心介绍了具身智能赛道新晋百亿独角兽一目科技的基本情况,以及触觉传感器赛道的发展现状,核心干货如下:

1. 当前人形机器人行业核心痛点:现有机器人能完成跑跳等大动作,但精细操作能力差,核心缺触觉感知能力,这也是触觉赛道成为资本风口的核心原因。

2. 一目科技的核心技术突破:针对传统方案体积过大的痛点,将视触觉传感器厚度压缩至小于3毫米,是目前公开同类量产产品的领先水平,同时解决了柔性材料耐用性、算法温漂和感知盲区问题,90%核心触觉指标达到类人水准,实现从材料到算法全链路自研。

3. 赛道增长数据:2026年国内人形机器人专用触觉传感器市场规模预计突破3.2亿美元,同比增速超160%,为所有零部件中增速最高的细分赛道,全球市场2031年预计突破350亿美元,属于高速增长的新赛道。

本文给具身智能、机器人相关品牌商展现了当前赛道的技术、消费趋势,以及研发和品牌建设的参考方向,核心干货如下:

1. 行业与消费趋势:当前具身智能的发展痛点已经从运动能力落地转到精细操作能力升级,触觉感知成为市场核心刚需,赛道增速极快,B端和C端对机器人实用能力的要求不断提升,倒逼感知零部件升级,是新的重要增长点。

2. 产品研发参考:行业普遍忽略体积痛点,一目科技从小切口切入实现突破,选择走分工开放的基础设施路线,避开和头部大厂整机业务的直接竞争,错位发展的思路值得品牌参考。

3. 品牌与商业化参考:一目科技采用可拆可合的三层商业模式,可满足不同客户的搭配需求,还通过推动开源数据集项目抢先建立行业标准,帮助抢占生态位,对品牌长期建设有较大参考价值。

本文给机器人相关赛道卖家梳理了具身智能赛道的新增量机会、风险与可学习经验,核心干货如下:

1. 增量市场机会:触觉传感器是当前人形机器人所有零部件中增速最高的细分赛道,2026年国内市场规模同比增速超160%,全球市场2031年将突破350亿美元,市场处于急速扩容阶段,存在大量未被满足的增量需求。

2. 赛道布局方向参考:当前行业最大卡点是前端感知能力不足,没有高质量触觉数据就无法训练出合格的具身智能模型,卖家可选择切入基础设施层,走开放分工路线,无需强求布局整机业务,做细分核心部件反而能抢占高价值生态位。

3. 风险提示:当前触觉技术整体还处于发展初期,虽有行业共识、技术路线收敛、成本可下降三重确定性,但技术仍未完全成熟,需要持续投入底层技术研发,提前布局数据标准才能建立足够壁垒。

本文给机器人相关制造工厂梳理了触觉赛道的产品需求变化、商业机会与数字化转型的启示,核心干货如下:

1. 产品生产与设计需求:当前下游客户对触觉传感器的核心需求集中在三点:足够小的体积,厚度要能适配小型机器人手部不挤占操作空间;材料要同时满足柔性可感知形变和工业场景耐用要求;算法要能实现高精度解算,解决温漂和感知盲区问题,这些都是生产端需要匹配的设计方向。

2. 新增商业机会:当前人形机器人赛道进入爆发期,触觉传感器是核心卡脖子零部件,市场需求急速增长,有技术积累的工厂可以切入该细分赛道,为各类机器人本体企业、灵巧手厂商做配套,获得大量增量订单。

3. 数字化转型启示:工厂可以参考一目科技的路线,聚焦底层技术自研,从核心材料到芯片都自主研发积累知识产权,同时参与行业数据标准建设,抓住物理AI发展的机遇,推进自身技术的数字化升级,抢占产业先机。

本文给具身智能赛道相关服务商梳理了行业发展趋势、客户核心痛点与可行解决方案方向,核心干货如下:

1. 行业发展趋势:具身智能行业的关注点已经从大模型研发、人形机器人本体制造,转向解决核心痛点触觉感知,2026年被行业称为触觉元年,已经获得资本和产业端的双重共识,视触觉方案已经逐渐收敛为主流技术路线,未来市场空间十分广阔。

2. 客户核心痛点:当前全行业的核心痛点是机器人触觉感知能力弱,市面上缺少体积足够小、性能稳定可靠的量产触觉传感器,同时物理世界的高保真触觉数据资产几乎空白,严重制约了具身智能大模型的训练迭代。

3. 解决方案参考:服务商可参考一目科技的全栈自研路线,从材料、芯片、算法三个维度同时优化,推出符合市场需求的高性能产品,也可以参与共建开源触觉数据集,参与建立行业标准,为全行业提供标准化的触觉数据服务,切入基础设施赛道抢占先机。

本文给机器人相关平台商梳理了当前行业的核心需求、新赛道布局方向与风险规避思路,核心干货如下:

1. 行业对平台的核心需求:当前具身智能行业走向开放分工,不同企业专注不同环节,迫切需要平台对接传感器厂商、整机厂商、算法厂商,同时行业缺少统一的触觉数据标准和接口,需要平台牵头推动标准建立,打通行业链路。

2. 平台招商与布局机会:触觉传感器是当前人形机器人赛道增速最快的细分领域,已经跑出估值百亿的独角兽一目科技,大量优质创业项目正在涌现,平台可以重点布局该细分赛道,引进相关优质项目入驻,打造平台新的增长曲线。

3. 风向规避思路:触觉赛道当前技术迭代快,尚未完全成熟,平台需要优先筛选技术壁垒高、全栈自研、有大量知识产权积累的项目,规避技术不成熟项目的风险,同时可以帮助项目对接下游客户,推动量产落地,助力赛道发展。

本文给人工智能产业研究者提供了具身智能赛道的最新产业动向、待研究的新问题与新商业模式参考,核心干货如下:

1. 最新产业动向:当前中美在物理AI领域基本处于同一起跑线,差距极小,具智能发展的核心卡点已经从大模型研发、运动能力实现,转到前端感知层升级,触觉感知已经成为行业共识,2026年被称为触觉元年,赛道融资升温,一目科技成为新晋百亿估值独角兽,资本已经大规模入局触觉赛道。

2. 行业待研究的新问题:当前行业存在三个层层递进的核心卡点:感知层物理世界数据采集能力不足、操作执行层类人精准操作能力不够、大脑模型层缺乏高质量训练数据,其中感知层是最核心的短板,物理世界高保真触觉数据的缺失,是当前行业亟待解决的新问题。

3. 新商业模式参考:一目科技采用可拆可合的三层商业模式,走开放分工的基础设施路线,打造从硬件到数据集再到模型赋能的全链路,还推动开源数据集建立行业标准,抢占物理AI数据守门人的生态位,这种差异化的商业模式值得深入研究。

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 outlines the basics of Yimu Technology, a new $10 billion unicorn in the embodied intelligence track, and presents the current development status of the tactile sensor sector. Key takeaways are as follows:

1. Core pain point of the current humanoid robot industry: Existing robots can perform large movements such as running and jumping, but lack fine operation capabilities due to the absence of mature tactile perception. This gap is the core reason tactile sensing has become a capital hotspot.

2. Yimu Technology's core technical breakthrough: Addressing the oversize issue of traditional solutions, the company has compressed the thickness of its vision-tactile sensor to less than 3 millimeters, a leading level among mass-produced comparable products publicly disclosed. It has also solved problems including durability of flexible materials, algorithm temperature drift and perception blind spots, with 90% of core tactile indicators reaching human-like levels, achieving full in-house research and development from materials to algorithms.

3. Track growth data: China's market size of dedicated tactile sensors for humanoid robots is projected to exceed $320 million by 2026, with a year-on-year growth rate of over 160%—the highest among all robot component segments. The global market is expected to surpass $35 billion by 2031, making this a fast-growing new track.

This article shares insights on industry technology, consumer trends, and reference directions for R&D and brand building for brands active in embodied intelligence and robotics. Key takeaways are as follows:

1. Industry and consumer trends: The core development bottleneck of embodied intelligence has shifted from realizing locomotion capability to upgrading fine operation, making tactile perception a core market demand. The track is growing extremely rapidly, and rising requirements for practical robot performance from both B-end and C-end customers are forcing upgrades to perception components, creating an important new growth opportunity.

2. Product R&D reference: While the industry generally overlooks the size pain point, Yimu Technology achieved a breakthrough by focusing on this overlooked niche. It pursues an open,分工-focused infrastructure strategy that avoids direct competition with large tech giants' complete robot business, and this错位发展 approach is a useful reference for brands.

3. Brand and commercialization reference: Yimu Technology adopts a modular three-layer business model that can meet the matching needs of different customers. It also has built an early industry lead through promoting an open-source dataset project to establish industry standards, a strategy that offers significant reference value for long-term brand building.

This article sorts out new growth opportunities, risks and actionable lessons for sellers in robotics-related tracks. Key takeaways are as follows:

1. Incremental market opportunity: Tactile sensors are the fastest-growing segment among all humanoid robot components, with China's market projected to grow over 160% year-on-year by 2026 and the global market set to exceed $35 billion by 2031. The market is expanding rapidly, with large unmet incremental demand.

2. Track layout reference: The biggest bottleneck in the current industry is insufficient front-end perception capability—qualified embodied intelligence models cannot be trained without high-quality tactile data. Sellers can enter the infrastructure layer with an open,分工-focused strategy, instead of forcing entry into the complete robot business. Focusing on core niche components can help capture high-value ecosystem positions.

3. Risk warning: Tactile technology is still in an early development stage overall. While it benefits from three certainties—industry consensus, converging technology routes and foreseeable cost reduction—the technology is not yet fully mature. Sustained investment in underlying R&D and early layout of data standards are required to build sufficient barriers.

This article sorts out changing product demand, new business opportunities and digital transformation insights for manufacturing factories in robotics-related fields. Key takeaways are as follows:

1. Product manufacturing and design requirements: Downstream customers' core requirements for tactile sensors center on three points: small enough size and thickness to fit small robot hands without crowding operating space; materials that deliver both flexible deformation perception and durability for industrial scenarios; and algorithms that enable high-precision calculation to solve temperature drift and perception blind spot issues. All these are design directions manufacturing end needs to adapt to.

2. New business opportunities: As the humanoid robot track enters an explosive growth phase, tactile sensors are a core bottleneck component with rapidly surging market demand. Factories with existing technical accumulation can enter this niche segment to supply complete robot manufacturers and dexterous hand vendors, securing large incremental orders.

3. Digital transformation insights: Factories can follow Yimu Technology's example by focusing on independent underlying R&D, building intellectual property through in-house development from core materials to chips, and participating in the formulation of industry data standards. This allows them to seize the opportunities brought by physical AI, advance their own technical and digital upgrades, and capture an early lead in the industry.

This article sorts out industry development trends, core customer pain points and feasible solution directions for service providers in the embodied intelligence track. Key takeaways are as follows:

1. Industry development trends: The embodied intelligence industry's focus has shifted from large model R&D and humanoid robot body manufacturing to solving the core pain point of tactile perception. 2026 is widely recognized across capital and industry as the "Year of Tactile Sensing", and vision-tactile solutions have gradually converged as the mainstream technology route, with very broad future market space.

2. Core customer pain points: The entire industry currently suffers from weak robot tactile perception. There is a lack of mass-produced tactile sensors that are sufficiently small, stable and reliable, while high-fidelity tactile data assets from the physical world are almost nonexistent—seriously constraining the training and iteration of large embodied intelligence models.

3. Solution reference: Service providers can learn from Yimu Technology's full-stack in-house development approach, optimizing across materials, chips and algorithms simultaneously to launch high-performance products that meet market demand. They can also participate in co-building open-source tactile datasets, help establish industry standards, and provide standardized tactile data services for the entire industry to enter the infrastructure track and capture an early lead.

This article sorts out current core industry demand, new track layout directions and risk mitigation strategies for robotics-related platform operators. Key takeaways are as follows:

1. Core industry demand for platforms: As the embodied intelligence industry moves toward open division of labor with different companies focusing on different links, there is an urgent need for platforms to connect sensor manufacturers, complete robot vendors and algorithm developers. The industry also lacks unified tactile data standards and interfaces, requiring platforms to take the lead in standard setting and connect industry value chains.

2. Investment and layout opportunities for platforms: Tactile sensors are the fastest-growing segment in the current humanoid robot track, which has already produced Yimu Technology, a unicorn with a $10 billion valuation, and a large number of high-quality startups are emerging. Platforms can prioritize layout in this niche segment, attract relevant high-quality projects to settle in, and build a new growth curve for the platform.

3. Risk mitigation strategies: The tactile track is currently experiencing rapid technological iteration and is not yet fully mature. Platforms should prioritize projects with high technical barriers, full-stack in-house R&D and substantial intellectual property accumulation to avoid risks from immature technologies. They can also help projects connect with downstream customers, promote mass production and commercialization, and support the development of the track.

This article provides the latest industry trends, unaddressed research questions and new business model references for AI industry researchers. Key takeaways are as follows:

1. Latest industry trends: China and the US are currently on roughly the same starting line in physical AI research, with a very small gap. The core bottleneck for embodied intelligence development has shifted from large model R&D and locomotion capability realization to upgrading the front-end perception layer. Tactile perception has reached broad industry consensus, 2026 is labeled the "Year of Tactile Sensing", sector financing is heating up, Yimu Technology has become a newly minted $10 billion valuation unicorn, and capital has entered the tactile track on a large scale.

2. New unaddressed questions for the industry: The industry currently faces three layered core bottlenecks: insufficient physical world data collection capability at the perception layer, insufficient human-like precision operation capability at the execution layer, and a lack of high-quality training data at the model layer. The perception layer is the most critical short board, and the lack of high-fidelity physical world tactile data is an urgent new problem to be solved for the industry.

3. New business model reference: Yimu Technology adopts a modular three-layer business model, follows an open,分工-focused infrastructure strategy, builds a full value chain from hardware to datasets to model empowerment, and promotes open-source datasets to set industry standards, taking the ecosystem position of "gatekeeper of physical AI data". This differentiated business model is worthy of in-depth research.

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.

答案,或许就藏在机器人动手能力的关键堵点上。

具身赛道又迎来一笔重磅融资。

近日,一目科技宣布完成E轮超10亿元融资,投后估值突破100亿元,跻身具身智能百亿估值俱乐部。多家产业方、一线人民币基金及全球化美元机构联合押注。这笔钱将用于产线订单的规模化量产交付,以及材料、芯片、算法和数据等底层技术的持续迭代。

消息一出,几乎所有人都想问同一个问题:这家公司凭什么值百亿?答案,或许就藏在机器人动手能力的关键堵点上。

一个残酷的现实是:今天的人形机器人能跑能跳、能打拳跳舞,情绪价值拉满,可一旦涉及精细操作便立刻露怯。拿扣扣子来说,这个对人类再简单不过的动作,对机器人却是巨大挑战。扣子的滑动、布料的肌理,仅凭视觉几乎无法判定。而背后缺失的,正是人类最核心的感知能力:触觉。

正是这个痛点的放大,把触觉传感器推向了资本的风口浪尖。从年初到年中,我们可以一窥融资新闻的关键词的变化:上半年还在谈“大模型”,后来变成“人形机器人”,到了下半年,“触觉”和“感知”开始频繁出现。

我们再看一下整体规模:2026年国内人形机器人专用触觉传感器市场规模预计突破3.2亿美元,同比增速超160%,是所有零部件中增速最高的细分赛道;放眼全球,2024年市场规模已超150亿美元,预计2031年将突破350亿美元。

在这样一个急速膨胀的市场,一目科技更像一家稍显“安静”的宝藏公司,扎扎实实做基础科研,像极了淘金热里的卖铲人,用心打造一把最趁手的“铲子”,只不过这把“铲子”,瞄准的是物理AI时代最稀缺的触觉感知力。

把人类的“五感”赋予机器人,90%触觉核心指标达类人水准

一目科技的故事,始于创始人兼CEO李智强的一个技术愿景:把人类的“五感”赋予机器。

创始人李智强拥有卡内基梅隆大学环境工程博士学位,研研究方向涉及微流控生物传感器和光谱芯片。据公开资料,其博士期间曾获得全美化学年会最佳学术报告奖。

进入具身智能赛道后,团队花了数月跑客户,从灵巧手厂商到机器人本体企业,再到夹爪供应商。一圈下来,客户反馈最集中的问题只有一个:体积太大。

传统视触觉方案厚度普遍超过20毫米,很难装进小型机器人手部,就算适配主流灵巧手,也会挤占操作空间、影响动作灵活性,很多场景根本用不了。所以做一款足够薄、能装进去的视触觉传感器,成了一目科技的突破口。

什么是视触觉?简单来说,是在柔性材料背面装一组微型光学系统,当传感器接触物体时,材料表面发生形变,摄像头拍下形变产生的光影变化,再通过AI算法解算出压力、纹理、滑动趋势等信息。形象一点说,就像给机器人指尖装了一台用光来“看”触感的显微镜。相比传统的压阻、电容方案,视触觉能捕捉到的信息密度高得多,接近人类指尖的感知水平,但体积问题一直是它大规模应用的最大障碍。

一目科技从几何光学设计、模型解算到芯片化技术,一路优化下来,把传感器厚度做到了行业同类产品的一半。他们的仿生视触觉传感器采用仿人体指腹形态,尺寸和厚度都很接近人类指尖。关键技术是一目自研的超表面硅光感光芯片,芯片能力决定了体积压缩的极限。据一目科技透露,其传感器厚度已迭代至小于3毫米。公司方面称,这一数据在目前已公开的同类量产产品中处于领先水平。

除了体积,视触觉传感器的难,还卡另外两个地方:材料和算法。

材料上要软,才能感知细微形变;又要耐用,才能在工业场景里扛得住反复按压。这两种特性理论上天然对立。但一目科技自己组建团队研发柔性电子皮肤材料,从高分子聚合物的单体开始研究,探索官能团的自恢复性和骨架结构的刚性。目前已经实现百万次按压、数万次横向剪切力下性能不衰减。

第三是算法。 传感器通过内置摄像头捕捉弹性材料接触物体时的形变,拍到的是“触觉照片”序列,再通过AI解算出压力、纹理、滑动趋势等信息。感知性能上,传感器能做到微米级形变解算、0.005N力分辨率、最高120fps输出帧率。针对传统方案容易出现的温漂和感知盲区,一目科技用自研算法做了优化,实现全解算无盲区,防护等级达到IP65。

综合来看,一目科技的90%触觉核心指标达类人水准。以上三个维度整合在一起,也构成了一目科技如今的全栈能力,而且从底层芯片到感知算法,一目科技实现了全部自研。截至目前,一目科技在全球布局了超过700件知识产权,拿了国家高新技术企业认定和省专精特新中小企业认定,通过了ISO三体系认证,四年内拿了12项国际大奖。

百亿赌注,押在了机器人的“手”上

“在大模型领域,中国是一个追赶者的角色。但是在物理AI领域,中美基本上在同一个起跑线上,差距几乎没有。”在刚刚开幕的WAIC 2026的访谈中,一目科技创始人李智强这样判断。

这话听起来颇有底气,但底气究竟从何而来?答案很朴素,甚至有些出人意料:大家同时被同一个麻烦绊住了:机器人能跑能跳了,但干不了细活。正是这个被无数人忽略的“细活”卡点,撑起了一目科技的百亿估值。

仔细拆解,这笔10亿融资投的远不止一家传感器公司。它投的是一个关键共识:当具身智能的腿已经迈出去,手却还僵在半空,谁能让机器人真正“上手”,谁就卡住了下一个十年最值钱的生态位。

在李智强看来,行业目前有三种不同的打法。

第一种是纯模型算法类。英伟达的Cosmos 3、Physical Intelligence的π系列模型,这些公司相信要做通用的具身智能大脑,不太关注硬件,主要聚焦数据和算法。

第二种是软硬件一体。特斯拉的Optimus、Figure,还有中国的智元机器人,从本体到算法全部自己做。这类公司需要非常强的研发能力,把硬件研发、制造和软件算法模型统一起来。

第三种,是一目科技的路线。李智强说:“我们认为在这个时间点,具身智能要发展,分工和开放是一个非常重要的方式。”一目科技做的是基础设施层,把硬件和算法连接起来的中间层,打通从感知入口到数据飞轮再到赋能模型的整条链路。

为什么基础设施层至关重要?

李智强把物理AI的瓶颈拆成了三个卡点。第一个是感知层,机器人怎么采集物理世界的数据。第二个是操作执行层,怎么做到类人的精准操作。第三个是大脑层,而大脑的前提是数据。

三个卡点层层递进,缺一不可。但行业目前最大的问题出在最前端:感知能力太弱。没有精准的感知,就没有高质量的数据;没有高质量的数据,就无法训练出理解物理世界的大脑。这正是一目科技选择触觉作为突破口的逻辑。

这一判断与英伟达CEO黄仁勋在2026 NVIDIA GTC Taipei场合的表态有所呼应。在NVIDIA发布首套人形机器人参考设计时,触觉灵巧手首次成为官方参考架构的重要组成部分。

当然,有人会问:触觉还处在四五十分的水平,值这个价吗?李智强给出的理由很诚实,也很有说服力。2026年被称为“触觉元年”,绝非偶然。价值层面,行业已经达成高度共识;技术层面,视触觉路线正在收敛为主流方案;成本层面,视触觉是唯一能沿着摩尔定律持续降成本的触觉模态。这三重确定性叠加在一起,让触觉传感器从一个小众课题,变成了能承载百亿估值的资本锚点。

在工业场景里,这套逻辑正在被验证。拿电池包插拔组装工序来说,工人完成一道工序大约需要30秒,传统工业机器人还难以胜任。一目科技的带反馈方案第一次测试花了100秒,远远不如人工。但到了今年四月,时间缩短到50秒,现在已经降到15秒,比人工效率高了一倍。

“这就像AlphaGo在收敛前难以战胜人类,一旦突破便不可逆转。”李智强说。机器人通过强化学习,以learning-based方式学会这个技能,相当于把老师傅的手艺数字化了。它的优势在于可以24小时不间断训练,人一天最多干8小时。

商业模式上,一目科技的做法是“可拆可合”的三层结构。最底层是触觉硬件,传感器和带传感器的灵巧手,目前对收入贡献最大。中间层是触觉数据集和标准化能力。最上层是模型赋能,仿真能力、语义理解的编码器,帮客户更快达到AGI能力。客户可以自己去选择、去搭配自己的需求。

基础设施的野望,在数据层面体现得最为宏大。李智强指出,物理AI目前最大的短板是数据资产。语言大模型有海量文本,视觉有ImageNet,但物理世界的高保真触觉数据几乎一片空白。于是他正在推动一个名为TouchNet的开源数据集项目,试图为物理AI打造一个类似于ImageNet的标准化数据底座。

“这肯定不是一个人的力量能做到的,但前提条件是我们要先把标准、接口和整个链条规范下来、打通。然后我们就可以去形成这样的数据资产和数据飞轮。”这句话背后,意味着一目科技试图走一步定义行业规则的先手棋。谁掌握了数据集的标准和接口,谁就有可能成为物理AI时代的“数据守门人”,而这个地位一旦确立,百亿估值恐怕只是一个起跳点。

因此,一目科技给自己的最终定位是Physical AI基础设施提供商。这个定位的底层自信在于,无论未来具身大脑走哪条技术路线,要实现超越脚本的灵巧操作,都必须依赖真实的物理交互反馈。当大多数玩家还在争抢一个完整的机器人时,一目提供的,正是那一套标准化的物理交互数据解决方案,为各类“具身大脑”输出标准化、高保真的物理信号输入接口。

那么,谁在领跑下一代基础设施?这个问题现在还远未到有定论的时候。过去20年,数据中心建设解决了信息之间的流动、存储和计算。现在,物理AI世界需要的是对物理世界的感知和理解能力。在这个新赛道上,谁能率先构建起以触觉感知和多模态融合为底座的基础设施,谁就能拔得头筹。

注:文/黎曼,文章来源:投中网(公众号ID:China-Venture),本文为作者独立观点,不代表亿邦动力立场。

【版权提示】亿邦动力倡导尊重与保护知识产权。未经许可,任何人不得复制、转载、或以其他方式使用本网站的内容。如发现本站文章存在版权问题,烦请提供版权疑问、身份证明、版权证明、联系方式等发邮件至run@ebrun.com,我们将及时沟通与处理。

已点赞

+1

{{ praise_text }}

好文章,值得被更多人看到

FAQ回顾

一目科技是做什么的?

一目科技是Physical AI基础设施提供商,核心研发具身智能领域的仿生视触觉传感器,90%触觉核心指标达类人水准,目前已完成E轮超10亿元融资,投后估值突破100亿元,底层芯片到感知算法全部自研。

人形机器人触觉传感器赛道市场前景如何?

2026年国内人形机器人专用触觉传感器市场规模预计突破3.2亿美元,同比增速超160%,是所有零部件中增速最高的细分赛道;全球2024年市场规模已超150亿美元,预计2031年将突破350亿美元,增长空间广阔。

一目科技的视触觉传感器有哪些技术优势?

该传感器厚度小于3毫米,为公开同类量产产品领先水平;自研柔性电子皮肤材料可实现百万次按压、数万次横向剪切力下性能不衰减;算法实现全解算无盲区,防护等级达IP65,90%触觉核心指标达类人水准。

具身智能领域当前主要有哪些技术路线?

目前主要有三类路线:一是纯模型算法类,聚焦具身智能大脑的数据和算法研发;二是软硬件一体类,自主完成机器人本体到算法的全链路研发;三是基础设施层,打造连接硬件与算法的中间感知赋能链路。

广告

下一篇

征集报名《品牌数字化服务商资源手册》

《品牌数字化服务商资源手册》展示数字化服务商风采,亿邦全年零售大会通发,触达更多资源,马上加入!

马上入驻

实习生 月入12万

尽管该薪资针对的是清华姚班等少数学生,但仍然可以一窥AI人才的争抢火爆。AI造富浪潮席卷而来,这当中,人才成为最重要的王牌,往往能改变潮水的方向。

未来零售 AI

2026-07-20 04:36:38