• 亿邦公众号
    • APP下载

烧出来的头部 钝掉的赛道:具身智能的五重资本困局

作者:Judy 文章来源:IT桔子

2026-06-16 16:06

作者:Judy 文章来源:IT桔子

2026-06-16 16:06

邦小白快读

EN
全文速览

本文核心梳理了当前具身智能赛道的五大资本困局,点明赛道长期有价值但当下处于泡沫膨胀期,核心干货如下:

1. 当前行业资本分配极度不均衡,2026年上半年460亿元总融资中,超七成被前20家头部企业拿走,剩余200多家中小企业仅分享不到三成,种子轮加天使轮合计融资占比仅3%,草根创业者入场通道几乎关闭。

2. 当前的头部企业大多是资本堆出来的,极少有可量化的营收、出货量支撑,也就是产品-市场匹配度验证严重不足,估值脱离基本面,部分企业一个月内估值就能跳升3倍多,泡沫特征明显。

3. 如果是想入局的普通人,可以参考草根逆袭的路径:不卷热门融资概念,坚持核心技术全栈自研,先在消费端跑通规模化,用真实用户反馈迭代技术,避开烧钱换融资的内卷陷阱。

本文分析了当前具身智能赛道的发展现状,对布局相关业务的品牌商来说,核心干货如下:

1. 消费与行业趋势层面,当前具身智能还没有清晰的产品定位,用户需求不明确,商业模式未验证,技术路线也未收敛,盲目all in赛道的风险很高,赛道的关键商业化验证期在2028-2029年,可以等待节点验证后再加大投入。

2. 资本与竞争层面,赛道马太效应严重,资本扎堆头部企业,融资排名已经取代产品排名成为话语权核心,资本定价功能已经失效,品牌做营销宣传不要过度炒融资概念,要回归产品本身的竞争力。

3. 供应链层面,头部整机品牌都在推进纵向整合自研核心零部件,品牌如果布局整机赛道,提前规划核心部件自研能降低长期成本,避免后续被供应商卡脖子,也能提升自身的利润空间。

当前具身智能赛道长期价值确定,但当下泡沫特征明显,对想入局的卖家来说,核心干货信息如下:

1. 机会层面:具身智能是AI落地物理世界的必经方向,长期属于高增长赛道,创业机会主要集中在整机企业做不了或不愿意做的低优先级零部件领域,走真实商业化路径的草根项目也有突围空间。

2. 风险提示:当前赛道资本分配极度失衡,早期草根项目拿融资难度极大;拿到国资投资会附带地理锁定、对赌约束、退出限制三重枷锁,一旦商业化不及预期,这些承诺就会变成沉没成本,还会稀释创业团队的控制权。

3. 可参考的路径与应对:不要跟风走「烧钱做Demo吸引下一轮融资」的融资飞轮,优先走「真实销售获客-用户反馈迭代技术-降本扩销」的规模飞轮,密切关注2028-2029年的产业验证窗口,提前做好现金流规划应对可能的资本退潮。

当前具身智能赛道的资本热潮给工厂带来了相应的机会与启示,核心干货如下:

1. 产品生产与需求层面:当前赛道处于早期,整机企业对外部零部件有采购需求,但参考智能手机产业链的发展规律,等整机企业出货量起来、规模扩大后,一定会走核心部件自研的垂直整合路径,外采需求会持续萎缩。

2. 商业机会层面:只有布局整机企业做不了或者不愿意做的低优先级部件,工厂才有长期生存的空间,如果做热门核心零部件,后续很容易被整机企业自研替代,陷入价格战甚至被迫转型。

3. 扩张启示:当前地方国资大量投资具身智能项目,要求工厂落地本地,会拆散工厂原有供应链,推高管理成本,工厂不要盲目跟风拿国资扩产能,要避免重蹈光伏产业产能过剩、集体亏损的覆辙,产能调整要基于自身商业需求,而非投资绑定要求。

当前具身智能赛道处于泡沫膨胀期,衍生出大量行业痛点与服务需求,对服务商来说核心干货如下:

1. 行业发展趋势:2026年上半年赛道总融资超460亿元,资本热度极高,资金集中流向头部企业,国资参与了42%的大额融资,零部件赛道融资热度最高,当前赛道正在复刻自动驾驶的泡沫发展路径,2028-2029年将迎来商业化验证期,资本退潮风险较高。

2. 核心客户痛点:中小草根创业者早期融资难,入场通道几乎被资本堵塞;头部企业绑定国资投资后,普遍存在供应链分散、管理成本高、退出受限、控制权稀释的问题;独立零部件企业面临下游整机自研替代的风险,增长空间被持续挤压。

3. 可切入的解决方案:可以针对草根创业者搭建早期项目孵化、融资对接服务;针对头部企业开发多区域供应链管理优化服务;针对零部件企业拓展非整机领域的客户渠道,帮助其降低被替代的经营风险。

当前具身智能赛道的发展现状,对布局赛道的平台商来说有以下核心干货:

1. 赛道需求层面:当前赛道资本分配严重失衡,7成资金流向已经有充足弹药的头部企业,大量有技术的中小草根项目拿不到早期融资,早期项目的融资、孵化需求缺口极大,平台可以针对性开放早期项目绿色通道,挖掘潜力项目,建立差异化竞争优势。

2. 招商与运营层面:招商不要盲目追逐高估值的头部网红项目,要优先考察项目的真实营收、用户反馈和商业化闭环,警惕估值泡沫;运营层面可以多扶持走真实商业化路径的草根项目,丰富平台的项目生态。

3. 风险规避方面:要牢记2028-2029年是赛道商业化验证的关键节点,提前做好风险防控,对于绑定地方国资的项目,要提前明确退出规则,规避地理绑定带来的退出风险,同时持续跟踪四个拐点信号判断赛道风向,及时调整平台策略。

本文针对当前具身智能赛道的资本格局提出了新的观察,对产业研究来说核心干货如下:

1. 产业新动向:2026年上半年具身智能赛道融资总额超460亿元,共288起融资涉及226家企业,资本热度极高;数亿元以上的大额融资中,国资背景机构参与率达42%,和光伏早期的参与度高度相似;零部件赛道平均单家融资3.2亿元,为全产业链最高,整体赛道处于泡沫膨胀期。

2. 新发现的产业问题:当前赛道的马太效应本质是资本从众内卷,而非市场验证后的自然结果,不是产品越好越拿到更多钱,而是头部拿了钱所以更能拿到钱;此外还存在估值脱离基本面跳升、国资投资附带隐性枷锁、零部件卖水逻辑存在致命盲区、创新管道堵塞等五重困局,赛道已经出现资本钝化的症状。

3. 研究参考:本文提出了四个可跟踪的产业拐点信号,还对比了当前赛道两种截然不同的商业模式:头部企业的融资飞轮和草根项目的规模飞轮,为研究产业周期和商业模式提供了新的分析框架,也为预测资本退潮时点提供了参考依据。

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 five key capital dilemmas facing the current embodied intelligence sector, noting that while the track holds long-term value, it is currently in a period of expanding hype. Key takeaways are as follows:

1. Capital allocation in the sector is extremely uneven. Of the total RMB 46 billion in financing raised in the first half of 2026, over 70% went to the top 20 leading companies. The remaining 200+ small and medium-sized enterprises split less than 30% of total capital, and combined seed and angel round financing accounts for only 3% of the total. The entry path for grassroots entrepreneurs is all but closed.

2. Most current leading companies have been built on heavy capital injections, with very few backed by quantifiable revenue or shipment volumes. In other words, product-market fit remains largely unproven, valuations have disconnected from fundamentals, and some companies have seen their valuations jump more than three times in a single month — clear signs of hype and overvaluation.

3. For ordinary people looking to enter the space, a grassroots breakthrough path is recommended: avoid competing on overhyped fundraising concepts, stick to full-stack in-house development of core technology, first scale up operations in the consumer market, iterate technology based on real user feedback, and steer clear of the capital-intensive vicious cycle of burning cash to secure more funding.

This article analyzes the current state of the embodied intelligence sector, with key takeaways for brands looking to lay out related business as follows:

1. On consumer and industry trends: Embodied intelligence still lacks clear product positioning, well-defined user demand, a validated business model, or a converged technology path. Going all-in blindly carries high risk. The critical commercial validation window for the sector falls between 2028 and 2029, so brands can wait for this milestone before scaling up investment.

2. On capital and competition: Matthew effect is extremely pronounced in the sector, with capital clustering around leading companies. Financing ranking has replaced product strength as the core source of market influence, and the capital pricing function has failed. Brands should not overhype financing concepts in marketing, and should refocus on building core product competitiveness.

3. On supply chain: Leading complete-machine brands are all pushing for vertical integration and in-house development of core components. For brands entering the complete-machine market, planning in-house core component development early will cut long-term costs, avoid future being held hostage by suppliers, and increase own profit margins.

While the embodied intelligence sector has clear long-term value, it currently shows strong signs of hype. Here are key takeaways for sellers looking to enter the space:

1. Opportunity outlook: Embodied intelligence is an inevitable direction for AI to落地 the physical world, and will be a high-growth sector long term. The main entrepreneurial opportunities lie in low-priority component segments that complete-machine companies are unable or unwilling to address, and grassroots projects following a real commercialization path still have room to break through.

2. Risk warnings: Capital allocation in the current sector is extremely uneven, and early-stage grassroots projects face enormous difficulty raising funding. Taking investment from state-owned assets comes with three constraints: geographic locking, performance bets, and exit restrictions. If commercial performance falls short of expectations, these commitments will become sunk costs and dilute the founding team’s control.

3. Recommended path and strategies: Do not follow the crowd into the "fundraising flywheel" of burning cash on demos to attract the next round of financing. Prioritize the scaling flywheel of "acquire customers through real sales → iterate technology based on user feedback → reduce costs and expand sales". Keep a close eye on the 2028-2029 industry validation window, and plan cash flow early to prepare for a potential capital retreat.

The current capital boom in the embodied intelligence sector brings corresponding opportunities and implications for factories. Key takeaways are as follows:

1. On production and demand: The sector is still in an early stage, and complete-machine companies currently have outsourcing demand for external components. However, following the development pattern of the smartphone industry chain, once complete-machine companies ramp up shipment volumes and scale, they will inevitably shift to vertical integration with in-house development of core components, and external procurement demand will keep shrinking.

2. On commercial opportunities: Only by targeting low-priority components that complete-machine companies cannot or will not develop can factories secure long-term survival space. If factories enter popular core component segments, they will likely be replaced by in-house development from complete-machine makers later, and end up stuck in price wars or forced to pivot.

3. Implications for expansion: Large amounts of local state capital are currently flowing into embodied intelligence projects, requiring factories to locate locally. This requirement will break apart factories’ existing supply chains and drive up management costs. Factories should not blindly take state capital to expand capacity, and should avoid repeating the overcapacity and collective losses seen in the solar industry. Capacity adjustments should be based on the factory’s own commercial needs, not the requirements tied to investment.

The embodied intelligence sector is currently in a period of expanding hype, which has generated a large number of industry pain points and service demand. Key takeaways for service providers are as follows:

1. Industry development trends: Total financing in the sector exceeded RMB 46 billion in the first half of 2026, showing extremely high capital enthusiasm. Capital is concentrated in leading companies, and state capital participated in 42% of large financing deals. The components segment has seen the strongest financing enthusiasm. The sector is currently repeating the hype trajectory that autonomous driving went through, and will enter its commercial validation period in 2028-2029, carrying relatively high risk of a capital retreat.

2. Core customer pain points: Early-stage grassroots entrepreneurs face extreme difficulty raising funding, with entry channels all but blocked by concentrated capital. After tying up with state capital investment, leading companies commonly face problems including fragmented supply chains, high management costs, exit restrictions, and diluted control. Independent component companies face the risk of being replaced by downstream complete-machine in-house development, with their growth room constantly squeezed.

3. Actionable solution opportunities: Service providers can build early-stage project incubation and financing matchmaking services for grassroots entrepreneurs; develop multi-region supply chain management and optimization services for leading companies; and expand customer channels in non-complete-machine segments for component companies to help them reduce the operational risk of being replaced.

The current state of the embodied intelligence sector delivers the following key insights for platform players positioning themselves in the track:

1. On sector demand: Capital allocation is extremely uneven, with 70% of capital flowing to already well-capitalized leading companies. A large number of technically strong small grassroots projects cannot secure early-stage funding, creating a huge unmet demand for early-stage project financing and incubation. Platforms can open targeted green channels for early-stage projects to uncover high-potential projects and build differentiated competitive advantages.

2. On investment recruitment and operations: Do not blindly chase high-valuation high-profile leading projects when recruiting. Prioritize assessing projects’ real revenue, user feedback and commercial closed loop, and stay alert to valuation hype. On the operations side, platforms should support more grassroots projects following real commercialization paths to enrich the platform’s project ecosystem.

3. On risk mitigation: Remember that 2028-2029 is the critical milestone for sector commercial validation, and prepare risk control measures in advance. For projects tied to local state capital, clarify exit rules upfront to avoid exit risk stemming from geographic binding requirements. At the same time, continuously track four inflection point signals to gauge sector momentum and adjust platform strategies in a timely manner.

This article presents new observations on the current capital landscape of the embodied intelligence sector, with key insights for industrial research as follows:

1. New industry trends: Total financing in the embodied intelligence sector exceeded RMB 46 billion across 288 deals involving 226 companies in the first half of 2026, marking extremely high capital enthusiasm. Among large financing deals over RMB 100 million, state-backed institutions participated in 42% of deals — a share very similar to participation levels in the early stage of the solar industry. The average financing per company in the components segment is RMB 320 million, the highest across the industrial chain, and the overall sector is in a phase of expanding hype.

2. Newly identified industry problems: The Matthew effect in the current sector is essentially the result of herd mentality and vicious cycle among capital, rather than a natural outcome after market validation. It is not that better products get more funding, but rather that leading companies that have already raised capital are therefore better positioned to raise more capital. In addition, the sector faces five core dilemmas: valuations jumping far detached from fundamentals, implicit constraints tied to state capital investment, fatal blind spots in the "selling picks" logic for components, and blocked innovation pipelines. The sector is already showing symptoms of capital passivation.

3. Research references: This article proposes four trackable industry inflection point signals, and compares two very different business models in the current sector: the fundraising flywheel of leading companies and the scaling flywheel of grassroots projects. This provides a new analytical framework for studying industry cycles and business models, as well as a reference for forecasting the timing of a capital retreat.

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.

作者|Judy

来源 |IT桔子

昨天,我们的伙伴@钛媒体创投家,基于IT桔子关于具身智能产业链的融资数据,写了一篇很有深度的分析文章《2026年上半年,460亿砸向具身智能,只「养肥」20家企业?》

文章回答了五个问题:具身智能的钱都去哪了?谁在出钱?什么最值钱?哪里最激进?草根创业者入局还有机会吗?

这是一篇全面的客观解读。

读完之后,我在想一个问题:这些数字到底在说什么,数据又遮蔽了什么?

上半年就有460亿融资、226家企业拿到钱、前20家拿到七成——「马太效应」「资本壁垒」是当下格局的客观描述。这些词都对,但我想总结一下更后面的那层信息。

一、马太效应的底色:资本内卷

2026年上半年,具身智能赛道288起融资,披露总额超460亿元,涉及226家企业。

资金分配是极度失衡的——

前5家吸走约171亿(占37%),前20家合计拿走约330亿(超七成),剩余200多家公司分享约124亿,平均每家只有数千万。

比如千寻智能3个月累计融资45亿,相当于200多家公司总和的约三分之一。

最直觉的解读是「强者恒强的马太效应」,但我越想越觉得这个解读是不是有问题?

马太效应可以由两种力量驱动:

一种是市场验证后的正反馈——产品好、客户多、收入涨,资本追加;

另一种是资本惯性下的从众——头部拿了钱,所以头部更该拿钱。

当前具身智能的马太效应,后者的成分明显远重于前者。

有一个绕不开的事实是:在所有公开报道中,我们几乎找不到这些头部企业可量化的营收数据、出货量或客户数。

「头部」这个标签,更多是由融资金额定义的,靠产品-市场匹配(PMF)验证的少之又少。

融资排名、技术排名、商业排名,本是三件事,但在当前市场叙事中被悄然画上了等号。

「49家半年完成两轮及以上融资」这件事同样值得琢磨。

高频融资本身不说明企业强——它可能说明烧钱快、资本在抢份额,也可能说明估值锚不稳定。

那么,当资本的心态从「投明白了才加注」变成「怕错过了才抢进」,马太效应就变成了泡沫加速器。

二、估值加速度:

脱离基本面的跳升

有一家明星公司的估值变化让我印象深刻:

2026年1月,它A+轮估值30亿元,2月B轮估值100亿元——一个月内跳升3.3倍。

即便是当年高增长的SaaS公司,从A到B轮的估值倍增通常也需要6到12个月的产品迭代和收入增长支撑。

而现在的具身智能在一个月内发生的估值变化,实际只是「拿了更多钱」。

它自己有什么实际的技术突破吗、大额订单吗?

最近多家媒体都在报道,某灵巧手企业去年估值200亿,新一轮目标估值400亿+。这个数几乎追平宇树科技IPO发行估值420亿——而宇树是赛道公认的出货量第一。

凤凰网财经有一个统计揭示了一个更宏观的背离:

截至2026年一季度末,A股人形机器人概念板块总市值达11.89万亿元,同期全球人形机器人实际出货量仅约2000台。

行业龙头宇树科技市销率24.7倍,但73.6%的收入来自科研教育客户,工业刚需场景仅占约9%,2026年Q1营收增速从去年同期332.64%骤降至68.49%,扣非净利润同比腰斩;

追赶者云深处PS达41倍,2025年全年仅卖出1台人形机器人,对应PE约485倍。

估值到底锚在哪里?

当「增长预期赌博」取代「业绩验证」成为定价逻辑,估值就已经从信息的浓缩变成了想象的投影。

三、国资入场的隐性代价:

资本换落地,落地换枷锁

在《2026年上半年,460亿砸向具身智能,只「养肥」20家企业?》中,数据显示:数亿元以上的重磅交易中,国资背景投资机构参与率高达42%。

北京人形机器人创新中心A轮7亿,几乎被北京本土国资平台包揽;智平方拿了常州和成都的国资,在当地设第二总部或者区域研发中心的事儿没跑了。

「资本换落地、落地换订单」,是现在地方国资通过投资招商的标准模式。

说得再直白一些:国资投资附带地理锁定、对赌约束和退出限制三重枷锁。

被投企业必须在出资方所在城市落地产能、提供税收。

但是,对企业来说,供应链被拆散到多地,管理成本激增;产能调整从商业决策变成政治决策;IPO或并购时国资更关注「企业是否留在本地」。

退出收益反倒成了次要考量——对创业团队而言,这是隐性的控制权稀释啊。

光伏产业无疑是前车之鉴。

在「双碳」叙事下,地方政府一拥而上,2025年中国光伏名义产能超1100GW,全球需求仅约600GW,产能冗余近两倍;头部企业合计亏损超280亿元,77家上市光伏企业中近60%亏损,超150家破产清算;光伏与储能行业总负债达6.5万亿元。更触目惊心的是:多地政府以「以房入股」「代建厂房」等方式扶持的项目烂尾后,政府国资被起诉追债。

当「保姆式输血」变成「政企债务泥潭」,自由退出已经不可能。

具身智能赛道当前的国资参与度已与光伏早期高度相似。42%的大额融资参与率意味着近半数头部企业已绑定地理承诺,一旦商业化不及预期,这些承诺就变成了沉没成本。

四、「卖水人」的宿命:

零部件的安全边际并不安全

原文的数据显示,零部件赛道平均单家融资3.2亿元,是具身智能全产业链最高。

对零部件的公司投资逻辑看似无懈可击:无论哪家整机企业胜出,都离不开灵巧手、传感器、关节模组——「淘金热中卖铲子」。

但这个逻辑有一个致命盲区:整机企业有能力自研,也有动机自研。

智元机器人已出资或孵化了Fullive.AI、智身科技、智鼎机器人,在赛道内进行纵向供应链整合;宇树科技的核心关节模组也是自研的。这些企业出货量达到一定规模,自研的边际成本将低于外采——这是制造业的铁律。

零部件企业面临的更残酷的演化路径是:早期整机企业依赖外采,规模起量后开始自研,外采需求萎缩,零部件企业被迫降价或转型。

智能手机产业链已完整走过这条路了——触控芯片、屏幕模组、摄像头模组,早期都是独立供应商的天下,最终绝大多数被整机厂垂直整合。

零部件企业唯一的护城河是「整机企业做不了或不愿做的低优先级部件」。但「不愿做」和「不能做」是两回事,当市场足够大,整机企业随时可能改变策略。

「卖水人」安全边际的前提是淘金者永远需要外购铲子,但在具身智能这个高资本密度的赛道,整机企业有足够的资金和动机自己造铲子。

五、自动驾驶的镜鉴

IT桔子数据显示,2021年,中国自动驾驶赛道年融资近千亿元,创历史高点;2022年骤降跌幅74%。

毫末智行是这场泡沫最完整的标本:

2021年A轮近10亿、估值突破10亿美元跻身独角兽;

2025年现金流全面断裂,账户冻结至无力支付3.15万元执行标的,全员停工。

从独角兽到停摆,不过四年。

自动驾驶的泡沫路径极其清晰:资本狂热 → 头部吸金 → 估值脱离产品 → 商业化不及预期 → 资本退潮 → 出清。

当前具身智能赛道,正在精确复刻前三个阶段。

但具身智能面临的挑战比自动驾驶更严峻。

自动驾驶至少有一个清晰的PMF——Robotaxi,用户需求明确,商业模式清晰,技术路线收敛。具身智能三者皆无:PMF未定义,商业模式未验证,技术路线未收敛(VLA还是VLM?双足还是轮式?通用人形还是场景专用?)。

成本下降也不等于商业闭环成立。当前机器人在工业场景的核心任务仍是料箱搬运、上下料等重复性工作,距离「替代熟练工」还有巨大鸿沟。

假设一台17万元的人形机器人替代年成本8-12万元的工人,粗略回本周期8-14个月——但这还不含维护、换产调试、场景适配等隐性成本,计入后回本周期可能延长至18-24个月。

对大多数中小企业而言,这个ROI仍不足以支撑规模化采购决策。

泡沫不可怕,钝化才可怕

见证过无数风口的起落之后,我仍然觉得泡沫本身并不可怕。

相反,每一个重大技术革命都伴随泡沫——铁路、互联网、新能源汽车、AI大模型,概莫能外。

泡沫是技术革命的必要代价,它为早期高成本的技术探索提供了资本缓冲。

真正值得警惕的,是泡沫是否「钝化」——资本是否还在向有效的方向流动。

当前赛道的钝化症状已经很明显。

先看资金。七成钱流入20家已有充足弹药的企业,边际效用递减。一家已有45亿现金的公司再拿15亿,和一家有技术但缺5000万的中小企业拿到这笔钱,哪个对产业进步贡献更大?答案不言自明,但市场正在选择前者。

再看信号。估值脱离产品进度,融资排名取代产品排名成为行业话语权的主变量。当「谁拿的钱多」比「谁的产品更好用」更重要,资本的定价功能就失效了。

最让我不安的是创新这根管道。

种子轮加天使轮合计不到13亿元,仅占全赛道3%。草根创业者的入场通道几乎关闭——没有大厂背景、没有学术头衔,连天使轮都拿不到。

但赛道里并非没有草根逆袭的可能。

尤其是最近,刚好在流行2026年草根逆袭叙事——张雪机车成立仅两年多,在今年的世界超级摩托车锦标赛上,目前已经狂揽6座分站冠军,终结杜卡迪、雅马哈、川崎等豪门垄断。

它的路径很朴素——不卷最热的概念,核心技术全栈自研,先在消费端跑通规模化,用真实用户反馈迭代技术。

这类「规模飞轮」——真实产品销售获取用户和数据,驱动技术迭代,降低成本,扩大销售——和头部企业的「融资飞轮」——烧钱做Demo,吸引下轮融资,做更大Demo——形成鲜明对比。

但当前3%的早期资金占比,意味着这类草根突围者在起步阶段就要面对资本的厚墙。

最后是退出。42%的国资参与率意味着近半数头部企业已绑定地理承诺。一旦商业化不及预期,既无法收缩战线,也难以被并购,更不可能自由迁移。

判断与坐标

我认为当前具身智能赛道处于泡沫膨胀期。

这不意味着赛道没有未来——具身智能是AI走向物理世界的必经之路,长期价值毋庸置疑。但「方向正确」和「节奏正确」是两件事。

泡沫膨胀期的典型特征是:融资增速远快于商业化增速、估值增速远快于营收增速、产能规划远快于需求验证。

三条在当前赛道全部成立。

历史给出的时间窗口并不宽裕。

自动驾驶从狂热到退潮用了约3年,光伏从爆发到过剩用了约4年。具身智能的「量产元年」是2026年,按照类似节奏,2028-2029年将是验证期窗口关闭的关键时点。届时如果头部企业仍无法交出规模化的营收和可复制的商业模型,资本退潮将不可避免。

几个值得持续跟踪的拐点信号:

头部企业 估值/ 营收降至20倍以下(当前宇树24.7倍PS,远未达标);

国资参与率从42%降至25%以下(市场化的财务资本重新获得定价权);

整机企业自研零部件比例突破40%(零部件独立供应商的窗口期开始收窄);

第一起头部企业大幅裁员或估值下调事件(泡沫破裂的先行指标,通常是「最不该出事的那家出了事」)。

在此之前,所有关于「万亿市场」「十倍增长」的叙事,都应加一个前提注释:如果商业化能如期兑现。

而「如果」这两个字,恰恰是泡沫最脆弱的地方。

数据来源:钛媒体创投家、IT桔子、凤凰网财经、21财经、澎湃新闻等公开资料。

注:文/Judy,文章来源:IT桔子(公众号ID:itjuzi521),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

跨境电商资源对接

《全球跨境电商服务商资源手册》第二期手册现已发布,触达更多资源,了解跨境服务商信息,欢迎点击链接下载。

了解详情

对话麒盛科技黄小卫:智能床年入30亿背后20年的重与慢

麒盛科技以近乎偏执的克制,搁置风口、坚守医疗级底线,用二十年时间证明:比起追逐流量套利,把睡眠健康做成“难而正确”的事,盈利只是长期主义的自然结果。

跨境电商 马蹄社 马蹄动态

2026-06-16 15:50:22