• 亿邦公众号
    • APP下载

一年三轮 被雷军看好的具身公司又融资了

作者:张雪 文章来源:投中网

2026-06-29 09:45

作者:张雪 文章来源:投中网

2026-06-29 09:45

邦小白快读

EN
全文速览

本文核心是成立两年的具身智能企业纬钛机器人完成Pre-A轮融资,累计融资达数亿元,下一轮融资也接近完成,赛道资本关注度极高,核心干货信息如下:

1. 当前行业已经进入关键发展窗口期:AI从纯粹虚拟世界走向真实物理世界,具身智能、物理AI成为产业和资本共同追捧的新方向,全球科技巨头都已经将视触觉作为下一代具身智能的核心配置,行业告别纯视觉开环作业模式。

2. 纬钛主打视触觉核心技术,已经推出覆盖感知、执行、数据采集完整链路的四款新品,可给灵巧手提供类似人手的触觉感知能力,目前已经进入多个制造场景商业化落地,2025下半年以来订单暴涨,现有订单量已经远超去年全年总和。

3. 视触觉领域门槛极高,需要软硬件、算法和大量工程经验积累,靠堆钱堆人无法快速突破,不少头部企业自研失败后都选择和专业厂商合作。

本文披露了具身智能赛道的最新发展动态,对布局智能相关业务的品牌商有诸多参考干货,内容总结如下:

1. 产业与消费趋势清晰:AI已经从虚拟走向物理,具身智能进入落地窗口期,机器人要落地真实场景,触觉感知是绕不开的核心环节,全球科技巨头、独角兽都已经把视触觉纳入技术路线,行业共识已经形成。

2. 产品研发方向明确:现在市场最热的三大落地场景是自适应抓取、精准放置、柔性物体操作,这些场景都需要触觉反馈解决纯视觉无法处理的接触问题,品牌商做智能产品研发要重视视触觉的配置。

3. 商业化可走分工协同路线:不用盲目追求全栈自研,视触觉技术门槛高,工程积累周期长,核心环节交给专业供应商,反而能降低试错成本,更快推出落地产品。

4. 当前资本对硬科技的认知已经加深,不再只看概念,更看重客户验证、量产交付能力,品牌布局要以落地为核心。

本文透露了具身智能视触觉赛道的最新变化,给ToB科技领域卖家带来了清晰的机会与风险提示,干货总结如下:

1. 市场增长机会明确:从2025下半年开始,市场关注度和客户需求暴涨,纬钛当前在手订单已经远超去年全年总和,除了3C制造,汽车、物流、新能源、食品等多个领域都有落地需求,市场空间广阔。

2. 行业竞争风险提示:视触觉是非常系统化的能力,涉及硬件、软件、算法和大量工程经验,不是靠堆人堆钱就能快速做出来的,没有长期积累很难入场,不少头部企业自研都以失败告终,新入场卖家不要轻易碰没有积累的核心环节。

3. 可借鉴的商业模式:走专业化赋能路线,聚焦自己擅长的核心环节,给下游玩家做技术赋能,和行业玩家分工协同,而非全赛道竞争,比如纬钛就定位成触觉版的英伟达,只做触觉感知环节,和整机厂商合作共赢。

4. 经营节奏提示:不要被融资热潮和高估值带偏,公司主线永远是技术、产品和交付,硬科技创业要保持资金使用的纪律性,分阶段推进目标。

本文披露了智能制造升级的最新需求,给制造工厂带来了诸多商业机会和数字化升级的启示,干货总结如下:

1. 产品生产升级需求清晰:当前智能制造的精细操作环节,比如3C产品装配,单靠视觉无法解决接触瞬间的力度判断、形变检测、防滑检测等问题,必须搭配视触觉手眼协同能力才能解决,这是现有工厂自动化升级的核心痛点。

2. 存在明确的商业合作机会:现在头部机器人企业、多个领域的制造龙头都在寻找成熟的视触觉技术供应商,不少原本计划自研的头部企业都放弃自研转向对外合作,给有技术积累的供应商提供了大量合作机会。

3. 智能化升级启示:工厂推进数字化、智能化升级,不需要盲目追求全栈自研,核心技术环节可以和专业供应商合作,能有效降低试错成本,加快升级落地速度。

4. 赛道趋势利好国内制造工厂:当前中美具身智能发展路径不同,美国强在AI大脑,中国强在产业落地,国内工厂作为应用端,能率先吃到具身智能落地的红利,升级需求会持续释放。

本文梳理了当前具身智能赛道的最新发展趋势,给相关科技服务商提供了清晰的方向,干货总结如下:

1. 行业发展趋势明确:AI已经从虚拟软件领域走向真实物理世界,具身智能已经到了规模化落地的前夕,全球行业已经达成共识,机器人进入物理世界必须配备视触觉能力,整个行业彻底告别纯视觉开环作业模式,视触觉技术服务商的市场需求正在快速爆发。

2. 核心客户痛点清晰:下游机器人企业、智能制造工厂都缺乏成熟的视触觉感知能力,自研视触觉的技术门槛高、试错成本高,大部分企业很难做出来,对外采购成熟技术方案的需求非常旺盛。

3. 可行的解决方案路径:可以走专业化赋能的路线,聚焦视触觉的硬件研发、算法优化和工程积累,为下游不同场景的客户提供核心器件和技术支持,做细分领域的核心供应商,像纬钛一样定位成“触觉版的英伟达”,打开市场空间。

4. 当前客户已经从关注概念转向关注量产交付,技术落地能力才是服务商的核心竞争力,提前完成客户验证就能拿到更多机会。

本文披露了具身智能赛道的最新变化,对布局硬科技、机器人赛道的平台商有诸多参考干货,内容总结如下:

1. 赛道对平台的需求清晰:当前具身智能赛道玩家越来越多,但全栈自研的模式门槛太高、试错成本太高,大部分玩家都需要成熟的上游核心技术供应商赋能,平台需要搭建分工协同的生态,满足不同玩家的合作需求。

2. 平台招商方向明确:可以重点布局吸引像纬钛这样,在核心细分环节有长期技术积累、已经拿到头部客户验证的玩家,这类玩家不炒概念、落地能力强,资本认可度高,能带动整个赛道的生态成熟。

3. 运营风向规避提示:当前赛道热度很高,不少玩家盲目追求快融资、高估值,只讲故事不做交付,平台要引导玩家聚焦技术、产品和客户交付,避免泡沫放大,影响整个赛道的健康发展。

4. 接下来的机会明确:2026年是具身智能从产品化走向规模化落地的关键之年,行业比拼的是真实交付能力,平台可以围绕智能制造、柔性操作等落地场景搭建生态,促进上下游分工协同,抓住赛道升级的红利。

本文披露了具身智能赛道的最新产业动向,对产业研究者研究硬科技发展有很高的参考价值,干货总结如下:

1. 产业新动向清晰:当前AI发展已经进入从虚拟到物理的关键窗口期,全球已经形成新的技术共识,机器人要进入真实物理世界落地,必须配备视触觉感知能力,行业彻底告别纯视觉开环作业模式;中美形成了差异化的发展路径,美国强在AI大脑研发,中国强在产业端落地应用,优势在“产业手脚”环节。

2. 产业存在的新问题:当前赛道存在一定泡沫,不少参与者盲目追求快融资、高估值,偏离了技术和交付主线;全栈自研的模式浪费行业资源,视触觉领域的工程积累壁垒高,新参与者很难短时间跨越。

3. 出现了新的成熟商业模式:专业化分工赋能的商业模式逐渐清晰,上游核心环节的厂商聚焦自身擅长领域,给下游整机、场景玩家提供技术赋能,分工协同替代全栈自研,提升了整个行业的效率。

4. 对硬科技创业研究的启示:硬科技创业要以技术落地和客户交付为核心,要分阶段推进长期目标,不能被资本和估值牵着鼻子走,保持纪律性才能走得更远。

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 centers on Vita Robotics, a two-year-old embodied intelligence startup, which has just closed its Pre-A round of funding, bringing its total cumulative financing to hundreds of millions of yuan. Its next funding round is also nearly complete, highlighting the extremely high capital attention the embodied intelligence track is currently receiving. Key takeaways are as follows:

1. The industry has now entered a critical development window: As AI expands beyond the purely virtual world into the physical realm, embodied intelligence and physical AI have become the new focus for both industry players and investors. Global tech giants have all identified vision-touch sensing as a core component for next-generation embodied intelligence, and the industry is moving past pure-visual open-loop operation models.

2. Vita Robotics focuses on core vision-touch sensing technology and has launched four new products covering the full pipeline of perception, actuation, and data collection, enabling dexterous robot hands to achieve human-level tactile sensing capabilities. Its products have already been commercialized in multiple manufacturing scenarios, and orders have surged since the second half of 2025, with current order volume already far exceeding the total for the whole of last year.

3. The vision-touch sensing sector has extremely high barriers, requiring integrated software/hardware development, algorithm capabilities and extensive engineering experience. Simply scaling up investment and headcount cannot deliver rapid breakthroughs, and many leading companies that attempted in-house development have failed and turned to partnerships with specialized vendors.

This article discloses the latest developments in the embodied intelligence sector, offering valuable insights for brand owners looking to expand into intelligent business. Key takeaways are as follows:

1. Clear industrial and consumer trends: AI is moving from virtual to physical applications, and embodied intelligence has entered a window for large-scale commercialization. For robots to operate effectively in real-world scenarios, tactile sensing is an unavoidable core requirement. Global tech giants and unicorns have all added vision-touch sensing to their official technology roadmaps, marking a clear industry consensus.

2. Clear guidance for product R&D direction: The three hottest current commercial application scenarios are adaptive grasping, precise placement, and flexible object manipulation. All of these scenarios require tactile feedback to solve contact problems that pure vision cannot address. Brand owners developing intelligent products should prioritize the integration of vision-touch sensing capabilities.

3. Commercialization can follow a division-of-labor collaboration path: There is no need to blindly pursue full-stack in-house development. Vision-touch sensing has high technical barriers and requires long engineering accumulation, so outsourcing core links to specialized suppliers can reduce trial-and-error costs and speed up time-to-market for commercial products.

4. Capital has now deepened its understanding of hard technology, prioritizing customer validation and mass production delivery capabilities over pure concept hype. Brands should center their布局 on commercial落地.

This article reveals the latest shifts in the vision-touch sensing segment of the embodied intelligence track, outlining clear opportunities and risk alerts for B2B technology sellers. Key takeaways are as follows:

1. Clear market growth opportunities: Starting from the second half of 2025, market attention and customer demand have surged dramatically. Vita Robotics' current backlog of orders already far exceeds the full-year total of 2024, and demand exists across a wide range of sectors beyond 3C manufacturing, including automotive, logistics, new energy, and food, creating enormous market potential.

2. Industry competition risk alerts: Vision-touch sensing is a highly systematic capability that requires hardware, software, algorithms, and extensive engineering experience. It cannot be rapidly developed simply by scaling up investment and teams, and long-term accumulation is required to enter the market. Many leading companies have failed in their in-house development attempts, so new entrants should not easily enter core links where they have no existing accumulation.

3. A replicable business model: Pursue a specialized赋能 route, focus on your core areas of expertise, provide technical赋能 for downstream players, and collaborate via industry division of labor rather than competing across the entire track. For example, Vita Robotics positions itself as the "NVIDIA of touch", focusing solely on tactile sensing and building win-win partnerships with complete robot manufacturers.

4. Operational guidance: Don't get distracted by the financing boom and high valuations. A company's core focus should always be technology, product development, and delivery. Hard tech startups must maintain capital discipline and advance goals in phases.

This article discloses the latest demand for intelligent manufacturing upgrading, offering insights into commercial opportunities and digital transformation for manufacturing facilities. Key takeaways are as follows:

1. Clear demand for production upgrading: In the fine manipulation links required for modern intelligent manufacturing, such as 3C product assembly, pure vision alone cannot solve problems like force judgment at contact, deformation detection, and anti-slip detection. Only vision-touch hand-eye coordination can address these issues, which is the core pain point for current factory automation upgrading.

2. Clear commercial cooperation opportunities: Leading robot companies and manufacturing giants across multiple sectors are now searching for mature vision-touch sensing technology suppliers. Many leading firms that originally planned for in-house development have abandoned that approach and turned to external partnerships, creating massive cooperation opportunities for technologically accumulated suppliers.

3. Insights for intelligent upgrading: When advancing digital and intelligent transformation, factories do not need to blindly pursue full-stack in-house development. Partnering with specialized suppliers for core technology links can effectively reduce trial-and-error costs and accelerate the pace of upgrading implementation.

4. Track trends benefit Chinese manufacturing facilities: China and the U.S. follow different development paths for embodied intelligence: the U.S. holds an advantage in AI "brains", while China leads in industrial落地. As the core application end, domestic Chinese factories can capture the first-mover advantages of embodied intelligence commercialization, and upgrading demand will continue to be released.

This article outlines the latest development trends of the embodied intelligence track, providing clear strategic guidance for relevant technology service providers. Key takeaways are as follows:

1. Clear industry development trends: AI has expanded from virtual software into the real physical world, and embodied intelligence is on the eve of large-scale commercial落地. Global industry consensus has formed that robots entering the physical world must be equipped with vision-touch capabilities, and the industry has fully moved past pure-visual open-loop operation models. Market demand for vision-touch technology service providers is growing rapidly.

2. Clear core customer pain points: Downstream robot companies and intelligent manufacturing factories all lack mature vision-touch perception capabilities. In-house development of vision-touch technology has high technical barriers and trial-and-error costs, making it out of reach for most companies, so demand for purchasing mature technical solutions from external suppliers is extremely strong.

3. A proven solution path: You can adopt a specialized赋能 model, focus on vision-touch hardware R&D, algorithm optimization and engineering accumulation, provide core components and technical support for downstream customers across different scenarios, and act as a core supplier in your niche segment. Like Vita Robotics, you can position yourself as the "NVIDIA of touch" to unlock large market potential.

4. Customers have now shifted their focus from concept to mass production delivery, so technical落地 capability is a service provider's core competitiveness. Completing customer validation early will help you capture more market opportunities.

This article discloses the latest changes in the embodied intelligence track, offering valuable insights for platform operators布局 hard technology and robot sectors. Key takeaways are as follows:

1. Clear demand from the track for platforms: As more players enter the embodied intelligence space, the full-stack in-house development model carries extremely high barriers and trial-and-error costs. Most players require赋能 from mature upstream core technology suppliers, so platforms need to build an ecosystem of division of labor and collaboration to meet the cooperation needs of different players.

2. Clear investment and recruitment direction: Platforms should prioritize attracting players like Vita Robotics, which have long-term technical accumulation in core niche segments and have already secured validation from leading customers. These players avoid hype, have strong落地 capabilities, enjoy high capital recognition, and can drive the maturation of the entire track's ecosystem.

3. Risk guidance for operations: While the track is currently very hot, many players blindly pursue fast financing and high valuations, focusing only on storytelling rather than delivery. Platforms should guide players to prioritize technology, products and customer delivery to avoid expanding sector bubbles that harm the healthy development of the entire track.

4. Clear upcoming opportunities: 2026 will be a critical year for embodied intelligence to transition from productization to large-scale落地, and the sector will compete on the basis of actual delivery capabilities. Platforms can build ecosystems around commercial落地 scenarios such as intelligent manufacturing and flexible manipulation, promote upstream-downstream division of labor and collaboration, and capture the dividends of track upgrading.

This article discloses the latest industrial developments in the embodied intelligence track, offering high reference value for industry researchers studying hard technology development. Key takeaways are as follows:

1. Clear new industrial trends: AI development has now entered a critical window of expansion from virtual to physical domains, and a new global technical consensus has formed: for robots to落地 in real physical environments, they must be equipped with vision-touch perception capabilities, and the industry has fully moved past pure-visual open-loop operation models. China and the U.S. have formed differentiated development paths: the U.S. leads in AI "brain" R&D, while China excels in industrial end落地 applications, holding an advantage in the "industrial limbs" segment.

2. New emerging industrial problems: The sector currently carries a certain amount of泡沫, with many participants blindly pursuing fast financing and high valuations and deviating from the core focus on technology and delivery. The full-stack in-house development model wastes industry resources, and the engineering accumulation barrier in the vision-touch segment is so high that new entrants cannot cross it in a short time.

3. A new mature business model has emerged: A specialized division-of-labor赋能 business model has gradually clarified. Upstream core segment players focus on their areas of expertise, provide technical赋能 for downstream complete machine and scenario players, and collaborative division of labor has replaced full-stack in-house development, improving overall industry efficiency.

4. Insights for hard technology entrepreneurship research: Hard technology startups must center on technical落地 and customer delivery, advance long-term goals in phases, and avoid being led astray by capital and valuations. Maintaining discipline is required to build a sustainable long-term business.

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.

纬钛是小米布局具身的最早一环。

近日,成立两年的纬钛机器人(以下简称“纬钛”)宣布完成Pre-A轮融资,该轮融资由红鸟启航基金领投,领屹投资、永鑫方舟、韦豪创芯、宁波吉德等产业投资方联合参投,上海天使会联合投资,多家老股东资本持续加持。同时,下一轮融资也接近完成,累计融资额已达数亿元人民币。

此轮融资以产业资本为主,包括小米生态链企业以及制造业龙头企业相关基金,其中韦豪创芯背靠韦尔股份,在此之前一直专注于芯片领域,这是其在具身领域的首次出手。宁波吉德是小米生态链上重要企业,此次也是其在具身智能领域首次出手。

在去年,纬钛完成了由小米和祥峰分别领投的天使轮和天使+轮,更早之前的种子轮还拿到了AI领军人物的投资。

说起来,纬钛是小米布局具身的最早一环,其今年初还领投了自变量机器人,最近更是接连参与了物理AI公司章鱼动力(SynapX)的首轮融资与灵巧手企业曦诺未来的Pre-A轮融资,在具身智能产业链持续加码。

从软件AI到物理AI

“过去一年,具身智能和物理AI确实进入了一个重要的窗口期。”李瑞告诉投中网,“这个窗口期的本质是AI正在从纯粹的虚拟世界走向真实的物理世界。”

他进一步解释道:“前两年资本更多关注机器人的本体或‘大脑’,但随着机器人要真正去干活,大家意识到必须让它像人一样具备触觉和手眼协同能力。”

这种变化在资本市场和客户端都有直观体现。李瑞透露:“从2025年下半年至今的半年多时间里,我们感受到了空前的关注度提升和客户需求的暴涨。前面几个月拿到手的订单以及正在签的订单,已经远远超过去年全年总和。”

从产品矩阵来看,今年纬钛发布了四款视触觉新品,覆盖从感知到执行器再到数据采集的完整链路。

最受关注的是GF515——面向五指灵巧手的视触觉仿生指尖。这款产品在15mm宽的人造指尖空间内,实现了高分辨率触觉成像、高频率感知和多维度力信息获取。“没有触觉的灵巧手更像一只只会动的机械手,有了视触觉,灵巧手才真正开始具备精细操作能力。”李瑞形容道,“GF515的意义在于让灵巧手拥有了像人一样的手感。”

此外,GF220面向两指夹爪场景,Gripper 90是集成视触觉的夹爪执行器,UMI85则是视触觉数据采集设备,用于为具身智能模型提供真实物理交互数据。

在商业化方面,纬钛主要走两条主线。其一,进入真实的智能制造场景。比较典型就是与小米智能制造工厂的深度合作,将视触觉手眼协同能力应用于手机制造中的精细操作环节。“这类任务单靠视觉很难解决,真正的难点发生在接触的瞬间:有没有碰到、力度是否合适、有没有滑动、会不会压坏,这些都需要触觉反馈。”李瑞表示。

除小米外,纬钛还在3C、汽车、物流、新能源、食品等多个场景布局,与多家头部客户进行深度合作。

其二是与机器人企业的生态合作。 日前,某百亿估值灵巧手公司的最新一代产品也采用了纬钛的视触觉技术,该产品的指尖融合了视触觉传感器,最小可感知130微米的形变。“我们不是竞争关系,而是赋能关系。他们解决手的结构、自由度、驱动和执行能力,我们解决这只手接触物理世界时的触觉感知、力反馈和控制闭环。我们做的是触觉版的英伟达。”

对于现在的热门场景,李瑞归纳为了三大类:自适应抓取、精准放置、柔性物体操作。对此,他用了一个生动的比喻,“这三大类技能对应小孩子成长的三个阶段——从什么都可以抓,到学会拼乐高,再到会叠衣服做饭。”

规模化落地的关键之年

随着具身智能赛道升温,越来越多的创业者涌入视触觉领域。与此同时,一些融得巨资的具身智能头部企业也具备自研视触觉的能力。

谈及行业竞争,甚至是否被赶超时,李瑞的回答很笃定,“并不会,视触觉是一个非常系统化的能力,涉及硬件、软件、算法和大量工程know-how。即使是我们这样积累了十几年的团队,在做成产品的过程中依然踩了很多坑。这些坑不是靠堆人堆钱就能快速解决的。”

据透露,已有头部具身智能公司尝试自研视触觉后最终放弃,转而选择与纬钛合作。“在这个窗口期,更好的选择是分工协同,各自做擅长的事。”

创始团队的背景提供了底气。李瑞在MIT攻读博士期间与导师共同开创了机器人视触觉研究路线,发明了全球第一款分辨率可超越人类手指的视觉传感器。“我们从源头的创新到后续的持续迭代,一直在这个领域引领着发展方向。”李瑞说。

2026年,被纬钛定义为关键之年。

“如果说2024年是公司成立和技术产品化的起点,2025年是客户验证和商业化加速的一年,那么2026年将是纬钛从产品交付走向规模化落地的关键一年。”李瑞补充道,“2026年最重要的变化是,行业会从‘看谁讲得更大’进入到‘看谁交付得更实’。这对纬钛是机会,因为我们的技术本来就是为真实操作而生的。”

放眼全球,中美在具身智能领域呈现出不同的发展路径。美国强在AI大脑,中国强在产业手脚,而视触觉是连接智能和物理世界的神经末梢。

在纬钛看来,无论路线如何不同,一个共识正变得清晰:机器人要进入物理世界,必须有更好的感知和操作能力。“视觉很重要,但不能解决所有接触问题。在临场操作、柔性物体、精密装配、力控检测等场景里,触觉是绕不过去的。”

就最近而言,Meta、Google、Amazon等国际巨头已经将视触觉作为下一代具身智能的重要组成部分;同时,全球新晋具身独角兽Sharpa新一代灵巧手也全系标配了视觉触觉融合传感器,再看国内,日前,它石智航也重磅发布了OmniVTA视触觉世界模型。

可见,全球顶级物理AI产品均完成技术迭代,彻底告别纯视觉开环作业模式。

马斯克也曾指出,灵巧手工程量占整机开发的一半,而触觉传感器是其优化的重点方向。

拿钱要快,花钱要慢

回顾这一轮融资过程,纬钛创始人兼CEO李瑞感受最深的是投资人认知的快速变化。

“这轮融资过程中,很多投资人问得非常深入。”他回忆道,“他们不只是问市场空间和传感器参数,也会问很多跟量产相关的问题——客户验证、算法闭环、数据采集,以及未来和大模型结合的路径。”

这与两年前的情况形成鲜明对比。“两年前,我们跟很多投资人要进行科普,讲触觉是什么、能用在哪些方面。但在过去这一年半里,明显感觉到投资人已经对这个领域有一定概念,甚至进行了具身智能的整体分析后主动找上门。”

更让李瑞印象深刻的是,一些产业化投资人在说服他们拿钱时,会具体说明“我有什么样的场景,在这些场景里面可以用到你的传感器和算法,能够在我的产品里或产线上用上这个技术”。

从纬钛的感知来看,这说明资本对具身智能的理解在变深,不再只是看概念。越来越多的投资人认识到,视触觉不只是机器人上的一个传感器配件,而是物理AI时代感知、控制和数据闭环的入口。

关于这次融资的用途,纬钛计划重点投入三个方向,视触觉感知技术和核心器件的迭代、执行器落地和视触觉数采和模型闭环。“未来的机器人要具备泛化操作能力,必须从真实的接触数据中学习。”

具身智能赛道融资如烈火烹油,市场上更是频频出现“一轮未结束下一轮已敲定”的现象。

在这样的背景下,纬钛展现出了一份难得的清醒。李瑞坦言:“市场上确实有这种现象,也说明具身智能和物理AI赛道很热。纬钛在融资的过程中也明显感受到资本关注度很高,很多投资人希望尽早参与,也希望持续跟进。”

但他也强调:“我们内部一直比较清醒,融资不能变成公司的主线,公司的主线永远是技术、产品、客户和交付。融资的节奏太慢可能错过产业窗口,但融资的节奏太快,也可能让公司被估值牵着走。”

硬科技创业跟其他行业最明显的不同是,不是把估值抬得多高,而是每一轮融资之后,公司有没有能力用新的里程碑去支撑下一阶段的价值。

李瑞打了个比方,“融资只是加油,真正的比赛在工厂里,在产线上,在客户的手里。”同时,融资快不代表花钱快,估值高不代表可以犯低级错误。硬科技行业最怕的不是钱不够,而是钱让公司失去纪律性。

回顾两年来的整个创业历程,李瑞称:“通用机器人的目标有点像攀登珠峰,但我们是沿途下蛋。我们不是一上来就攻克最难的难题,而是有节奏、有分阶段地往这个目标迈进。未来真正强大的机器人一定不是单一技术路线的胜利,而是大脑、眼睛、手、触觉、控制和数据的系统融合。”

注:文/张雪,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

广告

下一篇

智谱GLM5.2逼近美国顶尖AI模型 多家海外企业切换使用

智谱GLM5.2开源模型性能逼近美国顶尖AI,成本仅为其五分之一,凭借高性价比和智能体任务优势,吸引多家海外企业切换使用,对西方AI实验室形成定价压力。

智能商业 AI 服务商 大模型

2026-06-29 09:38:12