广告
加载中

身体已就位、大脑却断粮:热钱狂涌下 谁在争夺机器人“灵魂”?

元君 2026-07-16 13:52
元君 2026/07/16 13:52

邦小白快读

EN
全文速览

本文核心讲当前人形机器人行业已经越过本体运动控制的门槛,核心竞争焦点转向负责感知决策的“大脑”也就是具身智能领域,一场核心技术卡位战已经开启。

1. 行业最新数据显示,2026年上半年国内具身智能赛道融资近440亿元,超过一半资金流向做“大脑”的企业,纯本体厂商仅拿到12.8%的融资,资本已经认准大脑是机器人产业的核心上限。

2. 当前行业核心瓶颈是高质量真实交互数据缺口超99%,现有机器人在开放场景下难以应对千变万化的需求,目前行业已经形成四路技术路线并进的格局,包括模型自研派、端侧算力派、软硬一体派、类脑架构派,多个路线都已经出现技术突破。

3. 未来机器人大脑会进化成全栈式智能中枢,最终催生出能在开放世界自主决策的智能机器人,产业发展空间极大。

当前人形机器人产业已经进入结构性调整阶段,品牌商可以把握产业变革带来的消费趋势和商业机会,提前布局核心竞争力。

1. 消费与产业趋势明确,机器人已经从工厂结构化场景走向家庭、商超、医院等开放C端场景,用户对机器人的自主决策能力要求越来越高,行业价值链重心已经从硬件本体向上游大脑迁移,大脑成为整条产业链中最具定价权的环节,掌握大脑技术才能掌握产品定价主动权。

2. 行业数据显示,2030年全球具身智能市场规模将达到2388亿美元,国内人形机器人销量将增长至27.12万台,市场增长空间巨大,同时政策已经将具身智能列为国家战略级产业,有明确的政策红利。

3. 当前开放场景机器人的智能能力还存在明显短板,品牌商可以结合自身场景优势,与技术厂商合作开发适配C端需求的智能机器人产品,抢占新赛道先机。

当前具身智能赛道处于高速增长期,卖家可以从中挖掘新的增长机会,同时提前规避潜在风险。

1. 赛道增长明确,2026年上半年国内具身智能融资近440亿元,接近2025年全年融资的80%,全年将刷新融资纪录,未来五年全球市场复合增长率达73%,属于确定性很高的增长赛道。机会层面,VLA模型、世界模型、数据基础设施、端侧芯片、开放场景解决方案都有大量资本和需求流入,To B、To C端都存在未被满足的需求。

2. 风险提示方面,本体赛道已经同质化加剧,定价权逐步流失,若选择本体加大脑的软硬一体路线,需要同时双线投入资金团队,现金流压力远高于单一赛道,必须平衡好融资节奏和产品落地进度。

3. 现在华为、腾讯、百度等科技巨头都下场开放了大脑相关的生态能力,卖家可以对接大厂生态,降低自身技术投入门槛,更快落地产品。

当前人形机器人产业的结构性变化,给工厂带来了新的生产需求和商业机会,也为工厂数字化升级提供了新方向。

1. 产品生产需求发生变化,随着机器人运动控制技术成熟,市场核心需求转向适配不同智能大脑的本体产品,面向家庭、商超、医疗等开放场景的人形机器人本体需求会持续增长,工厂可以提前调整产能布局,抓住新的订单机会。

2. 零部件生产领域的新机会,端侧部署已经成为机器人大脑的明确发展方向,机器人AI Box边缘计算设备、灵巧手、传感器、关节模组等核心零部件的量产需求持续提升,目前已经有头部厂商拿到规模化量产订单,给代工厂和零部件厂商带来稳定的业务增量。

3. 对工厂数字化升级的启示,工厂自身也可以引入具备自主决策能力的新一代机器人,解决传统工业机器人只能应对结构化场景的痛点,适配柔性生产、多品种小批量的生产需求,提升生产效率。

当前具身智能行业处于高速增长的爆发期,服务商可以挖掘新的客户痛点,布局对应的解决方案,抓住产业红利。

1. 行业发展趋势明确,人形机器人产业已经从本体竞争转向大脑竞争,价值链重心逐步向大脑层面迁移,大脑成为产业中最具定价权的生态位,未来市场规模会快速增长,2030年全球市场规模将达到2388亿美元,未来五年复合增长率73%,增长空间极大。

2. 客户核心痛点清晰,当前机器人大脑的核心痛点是合规可用的高质量真实交互数据缺口超过99%,绝大多数本体厂商不具备自主研发大脑的能力,开放场景落地也缺乏配套的技术支持,存在大量未被满足的服务需求。

3. 可布局的解决方案方向,服务商可以切入机器人数据基础设施赛道,积累合规交互数据,也可以对接大厂大模型和算力能力,为本体厂商提供定制化的大脑技术方案,或者针对零售、医疗等不同场景提供场景化落地部署服务,帮助客户快速推出产品。

具身智能产业的结构性变化,给平台带来了新的发展机会,平台可以调整布局策略,适配产业新需求。

1. 产业对平台的核心需求明确,机器人大脑研发需要海量算力、仿真环境和多模态大模型底座,中小创业公司没有能力独立搭建完整研发基础设施,同时大量创业公司也需要对接产业落地场景,需要平台提供生态对接服务。

2. 当前头部玩家已经探索出成熟的平台路径,华为推出CloudRobo平台接入昇腾算力矩阵,腾讯开源混元大模型为行业提供技术支持,科技巨头都选择做平台生态做大脑,发挥自身资源优势,避开硬件本体的价格战,这个路径值得平台参考。

3. 运营和风险规避要点,平台招商可以重点倾斜VLA模型、世界模型、数据基础设施、端侧芯片等大脑赛道企业,同时需要关注软硬一体派企业的现金流风险,帮助企业对接场景平衡融资与落地节奏,还要提前布局符合三域融合、端侧部署等未来技术方向的项目,规避技术路线迭代的风险。

当前人形机器人行业出现了清晰的结构性分水岭,产生了很多新的产业动向和新问题,具备很高的研究价值。

1. 产业新动向明确,目前人形机器人的运动控制“小脑”已经高度成熟,行业竞争焦点转向负责感知决策的“大脑”,2026年上半年国内具身智能赛道近440亿元融资中,超过一半流向大脑派企业,行业形成模型自研派、端侧算力派、软硬一体派、类脑架构派四路并进的格局,政策端已经将具身智能列为国家战略级产业,价值链重心正式向上游大脑迁移。

2. 行业出现多个待解决的新问题,核心问题是高质量机器人交互数据缺口超过99%,现有技术无法满足开放场景的需求,软硬一体路线的企业普遍面临较大的现金流压力,整个行业技术路线还未收敛,尚未跑通从模型到场景的成熟商业闭环。

3. 重点研究方向包括:不同技术路线的收敛趋势,数据基础设施的建设路径,端侧部署的工程优化方案,类脑架构对数据瓶颈的破解逻辑,以及如何出台配套政策推动产业健康发展,这些研究都能为产业发展提供重要参考。

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 explores the core inflection point in today's humanoid robot industry: the sector has now crossed the threshold of core motion control for robot bodies, and competition has shifted to "embodied intelligence"—the "brain" responsible for perception and decision-making, where a race for core technological positioning is already underway.

1. Updated industry data shows that in the first half of 2026, China's embodied intelligence sector raised nearly 44 billion yuan in funding. Over half of this capital went to companies developing robot "brains", while pure robot body manufacturers only secured 12.8% of total financing, confirming that investors view the brain as the core determinant of the industry's upper limit.

2. The industry's current core bottleneck is a shortage of over 99% of high-quality real-world interaction data, leaving existing robots unable to handle the diverse demands of open scenarios. Four competing technical approaches are currently being pursued in parallel: independent model development, edge-side computing power integration, integrated hardware-software solutions, and brain-inspired architecture, and all four paths have already seen meaningful technical breakthroughs.

3. Looking ahead, robot brains will evolve into full-stack intelligent hubs, eventually enabling fully autonomous decision-making in open unstructured environments, unlocking enormous industrial growth potential.

The humanoid robot industry is currently undergoing a period of structural adjustment. Brands can capitalize on new consumer trends and business opportunities brought by this industrial shift, and build core competitive advantages in advance.

1. Consumption and industrial trends are clear: robots are moving beyond structured factory environments to open consumer-facing scenarios such as homes, retail stores, and hospitals. Consumers are increasingly demanding autonomous decision-making capabilities from robots, and the center of the industry's value chain has shifted upstream from hardware bodies to robot brains. The brain has become the segment with the strongest pricing power across the value chain—only by mastering brain technology can brands take control of product pricing.

2. Industry data projects the global embodied intelligence market will reach $238.8 billion by 2030, while domestic humanoid robot sales in China will grow to 271,200 units, indicating enormous room for market growth. In addition, embodied intelligence has been designated a national strategic industry in China, with clear policy dividends available.

3. Open-scenario robots still have clear gaps in intelligent capability today. Brands can leverage their own scenario expertise to partner with technology firms and develop intelligent robot products tailored for consumer demand, to capture first-mover advantage in this new track.

The embodied intelligence track is currently in a period of rapid growth, offering sellers new opportunities for expansion while requiring proactive mitigation of potential risks.

1. Growth prospects are well-defined: in the first half of 2026, domestic embodied intelligence financing reached nearly 44 billion yuan, close to 80% of total 2025 full-year financing, and full-year 2026 figures are on track to hit a new record. The global market is projected to see a 73% compound annual growth rate over the next five years, making it a high-certainty growth sector. Areas including VLA models, world models, data infrastructure, edge-side chips and open-scenario solutions are seeing strong inflows of both capital and demand, with unmet needs in both B2B and B2C markets.

2. On the risk side, the robot body segment is already facing intensifying homogenization and eroding pricing power. Pursuing an integrated hardware-software route combining body and brain requires parallel investment in capital and teams, leading to far greater cash flow pressure than focusing on a single segment, so sellers must carefully balance financing pace and product commercialization progress.

3. Tech giants including Huawei, Tencent and Baidu have already opened up their brain-related ecosystem capabilities. Sellers can partner with these large firms' ecosystems to lower their technology investment barriers and bring products to market faster.

The ongoing structural shift in the humanoid robot industry has brought new production demand and business opportunities for factories, as well as a new direction for factory digital transformation.

1. Product demand is shifting: as robot motion control technology matures, core market demand has shifted to body products adapted for different intelligent brains. Demand for humanoid robot bodies targeted at open scenarios such as homes, retail and healthcare will continue growing, so factories can adjust capacity allocation ahead of time to capture new order opportunities.

2. New opportunities have emerged in components manufacturing: edge-side deployment has become a clear development direction for robot brains, driving sustained growth in mass production demand for core components including AI Box edge computing devices, dexterous hands, sensors, and joint modules. Leading players have already secured large-scale mass production orders, bringing stable business growth to foundries and component suppliers.

3. The shift also offers insights for factories' own digital upgrades: factories can introduce a new generation of autonomous decision-making robots to address the pain point of traditional industrial robots only being able to operate in structured scenarios. These new robots adapt to flexible production and high-mix low-volume manufacturing needs, helping to improve overall production efficiency.

The embodied intelligence industry is currently in a high-growth explosive phase. Service providers can identify new customer pain points, develop targeted solutions, and capture industrial dividends.

1. Industry development trends are clear: the humanoid robot industry has shifted from competing on robot bodies to competing on robot brains, with the value chain's center of gravity gradually moving upstream to the brain. The brain is now the industry's highest-value ecological niche with the strongest pricing power, and the market will grow rapidly: it is projected to reach $238.8 billion globally by 2030, with a 73% compound annual growth rate over the next five years, representing enormous growth potential.

2. Core customer pain points are well-defined: the biggest bottleneck for robot brains is a shortage of over 99% of compliant, high-quality real-world interaction data. The vast majority of robot body manufacturers lack the capability to develop brains independently, and also lack supporting technical support to deploy robots in open scenarios, leaving large unmet demand for services.

3. There are multiple promising solution areas for service providers to enter: they can build out robot data infrastructure and accumulate compliant interaction data; they can partner with large tech firms to access large models and computing power, and offer customized brain technology solutions for body manufacturers; or they can provide scenario-specific deployment services for retail, healthcare and other verticals to help clients launch products quickly.

Structural changes in the embodied intelligence industry have created new development opportunities for platforms, which can adjust their layout strategies to adapt to new industry demand.

1. Core industry demand for platforms is clear: robot brain development requires massive computing power, simulation environments and multi-modal large model base infrastructure, which small and medium-sized startups cannot build independently. In addition, many startups need support to access real-world industrial deployment scenarios, creating demand for platform-enabled ecological connectivity services.

2. Leading players have already proven mature platform models: Huawei launched its CloudRobo platform connected to its Ascend computing power matrix, Tencent open-sourced its Hunyuan large model to support the industry. Large technology firms are all focusing on building platform ecosystems for robot brains to leverage their existing resource advantages and avoid price competition in hardware bodies, a path that other platforms can reference.

3. For operations and risk mitigation: platforms can prioritize onboarding companies focused on brain-track segments such as VLA models, world models, data infrastructure and edge-side chips. They also need to monitor cash flow risks for integrated hardware-software companies, help connect these firms with scenarios to balance financing and commercialization progress, and proactively invest in projects aligned with future technology directions such as three-domain integration and edge-side deployment to mitigate risks from technical route iteration.

The humanoid robot industry is now facing a clear structural watershed, with many new industrial trends and open research questions that represent high research value.

1. New industrial trends are clear: the "cerebellum" of motion control for humanoid robots is now highly mature, and competition has shifted to the "brain" responsible for perception and decision-making. Of the nearly 44 billion yuan raised in China's embodied intelligence sector in the first half of 2026, over half went to brain-focused companies. The industry has formed a pattern of four parallel technical approaches: independent model development, edge-side computing power integration, integrated hardware-software solutions, and brain-inspired architecture. On the policy side, embodied intelligence has been designated a national strategic industry, marking the official shift of the value chain's center of gravity upstream to robot brains.

2. Multiple new unsolved problems have emerged in the industry. The core challenge is a shortage of over 99% of high-quality robot interaction data, and existing technology cannot meet the demands of open scenarios. Companies pursuing the integrated hardware-software route generally face significant cash flow pressure, technical routes across the industry have not yet converged, and a mature commercial loop from model to scenario has not yet been proven.

3. Key research directions include: the convergence trend of different technical routes, the construction path for data infrastructure, engineering optimization solutions for edge-side deployment, the logic of brain-inspired architecture for solving data bottlenecks, and how to design supporting policies to promote healthy industrial development. Research in these areas can provide important guidance for industrial development.

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.

当人形机器人的双腿能稳稳行走、双手可以灵巧抓握,行业却迎来一道清晰的分水岭。

数据显示,2026年上半年,近440亿元资金涌入国内具身智能赛道;与此同时,一个惊人的结构性变化也同时发生,超过一半的钱涌向了“大脑派”公司,而机器人本体厂商拿到的钱不到20%。

事实上,眼下的具身智能领域,运动控制这套“小脑”已高度成熟,负责理解世界、自主规划的“大脑”,仍被海量场景数据和技术瓶颈牢牢困住。

当身体不再是门槛,大脑便是一切上限。从VLA大模型到端侧算力芯片,从世界模型到数据基础设施,一场围绕机器人大脑的卡位战已经全面打响。

1

当“大脑”成为决胜点

今年5月,苏超联赛开赛前夕,一场别开生面的全自主机器人3V3足球对抗赛率先开打。现场,多台人形机器人追逐着一颗足球,它们没有遥控器、没有预设轨迹,全靠自身“大脑”实时感知场上局势、自主规划跑动路线,并决策传球还是射门。

虽然动作偶尔略显笨拙,但每一步都是机器自己“想”出来的。而在三年前,这样的场景还只存在于实验室的设想中。

彼时,行业还在为人形机器人能不能站稳、能不能走路而绞尽脑汁。但到了2026年,画风已经骤变。随着宇树、优必选、智元等厂商的产品已经能稳定行走、搬运物料,甚至完成拧瓶盖这类精细活,机器人“小脑”层面的运动控制已高度成熟。

不过,走得稳和想得明白之间,似乎还横亘着一道天堑。

要实现通用自主能力,一个合格的机器人大脑至少需要千万小时级别的高质量真实交互数据。而现实是,截至2026年初,全球合规可用的机器人数据只有50万小时,缺口超过99%。

数据的匮乏直接表现为大脑的“幼稚”。今天的机器人在结构化工厂里可以娴熟地装配零件,一旦走进家庭、商超这些开放场景,面对凌乱的桌面、突然闯入的宠物、千变万化的用户需求,它们却常常陷入窘境。

或许正是看到了这个致命短板,资本开始选择“用脚投票”。

据量子位不完全统计,2026年上半年(截至6月12日),国内具身智能赛道融资近438亿元,逼近2025年全年约554亿元的80%。按照这一速度,2026年全年将毫无悬念地刷新纪录。

但比总量更耐人寻味的是钱的流向。在近440亿元融资中,超过一半的资金涌向了“大脑派”公司;而纯“本体派”厂商只分到了约12.8%的份额,这个数字甚至低于生产灵巧手、传感器、关节模组等核心零部件企业14.4%的占比。

而大脑赛道内部的分化同样剧烈。VLA(视觉-语言-动作)模型一口吃掉了42%的资金,世界模型占27%,数据基础设施占20%,端侧芯片占11%。大量大脑派公司融资节奏快到平均一个月一轮,甚至有企业两轮融资间隔只有两周。

与此同时,科技巨头和云厂商们也纷纷下场。华为、腾讯、百度等均加码布局机器人大脑,生态战全面打响。

其中,华为推出CloudRobo平台,试图将机器人接入昇腾算力矩阵;腾讯则定位为产业生态中的“钛螺丝”,通过开源混元大模型解决机器人的思考问题。大厂集体选择做大脑而不做本体,除了有效避免在硬件本体上陷入价格战,更关键的是能发挥自身海量算力、仿真环境和多模态大模型底座的核心能力。

当身体的门槛被逐渐踏平,大脑的壁垒正越筑越高。随着具身智能竞赛规则的改写,这场关于“灵魂”的争夺才刚刚鸣枪。

2

谁在定义机器人大脑?

目前来看,具身智能大脑赛道已形成多路并进的格局。

其中,模型自研派以VLA与世界模型为核心壁垒,走模型定义大脑路线,是融资规模最大、估值增速最快的阵营。千寻智能是其中跑得最快的选手之一。

这家2024年1月才成立的公司,在2026年迎来了一轮令人瞠目的融资狂飙。其在2月连融两轮近20亿元,投资者包括云锋基金、红杉资本等顶级机构;3月再获10亿元融资,估值跃升至200亿元;随后又拿下15亿元,短短三个月累计吸金45亿元。

而让资本如此亢奋的,是它在技术上的一个标志性突破。今年1月,千寻智能的开源模型Spirit v1.5在性能上超越了美国头部公司Physical Intelligence的Pi0.5模型,成为中国首个实现此突破的开源模型。同时,其技术含金量也很快在零售、工业等产业端场景得到验证。

另一家走模型路线的无界动力,则押注隐空间世界模型与强化学习的双轮驱动路线。其逻辑是让机器人理解世界运行的物理规律,而非仅仅记忆场景。这意味着机器人可以走进任何一扇没见过的门,而不是只在熟悉的地方走路。

2026年6月,无界动力发布的MWA?具身通用大脑在斯坦福大学等机构联合发起的RoboCasa GR1 TableTop榜单上登顶全球第一,超越ACE-EGO-0、DIAL等多款主流模型。

端侧算力派则从大算力加大模型端侧部署切入,强调本地化实时决策与量产交付能力。这一阵营中,知行科技旗下的如一动力值得关注。

如一动力深耕机器人大脑领域,依托知行科技在自动驾驶领域积累的技术栈和工程量产经验,完成初期技术与工程验证。其核心产品机器人AI Box边缘计算设备,算力覆盖80、128至560TOPS,可推动具身智能端侧算法转向模型驱动。同时,结合专为端侧部署优化的VLN(视觉-语言-导航)与VLA(视觉-语言-动作)模型架构,机器人可摆脱对云端的绝对依赖,在本地高效完成实时感知、推理、决策等。

目前,如一动力的技术能力已在多个真实场景中得到验证。除了前文提到的足球对抗赛,在今年的北京亦庄人形机器人半程马拉松上,如一动力团队通过开发与部署自主导航避障算法,让机器人以不到两小时成绩成功完赛。其自研的AI BOX产品iRC100P(AI算力128TOPS)还获得国内头部具身智能企业规模化量产订单,并完成首批交付与部署。

此外,本体+大脑一体派同时拥有顶尖模型与自研本体,试图以软硬协同形成闭环。一体派的核心逻辑是同时掌控“身体”与“灵魂”,这一路线中,星海图是典型代表。

不过,这套逻辑听起来很性感,但代价也显而易见。因为要同时兼顾模型与本体,两条线都需要资金、团队,一体派企业的现金流压力通常会超过纯软件或纯硬件公司。因此,能否跑通这条路,也取决于企业是否可以在融资节奏和产品落地之间找到平衡点。

类脑架构派则试图绕开传统VLA的数据瓶颈,从仿生学底层重构大脑架构。智平方是这一方向的先行者。

2026年,这家公司发布了全球首个类脑具身智能系统NeuroVLA,借鉴了人类大脑皮层—小脑—脊髓的三级体系。不同于传统大模型对海量数据和算力的持续依赖,NeuroVLA依托人脑工作机制,可以让机器人通过更少的数据完成学习与决策。

当然,多路并进,殊途同归。

VLA派追求模型能力的极致突破,端侧派卡位算力部署的工程高地,一体派试图用软硬协同构建闭环,类脑派则从仿生学底层重构大脑架构。而所有玩家奔向的是同一个终点,让机器人真正理解世界。

未来,或许谁能在技术收敛之前完美跑通从模型到场景的商业闭环,谁就有望定义下一代机器人的思考方式。

3

下一站在哪里?

钱在往里涌,人在往前冲。但一个更底层的问题摆在那里:这个市场到底有多大?钱最终会流向价值链的哪个环节?

赛迪研究院《2026年未来产业十大赛道》报告指出,全球具身智能市场未来五年复合增长率73%,2030年市场规模有望达到2388亿美元。GGII也曾预测,2030年国内人形机器人市场规模将接近380亿元,销量将增长至27.12万台。

数字很大,但真正值得关注的不是规模,是结构。

过去两年,资本主要集中在硬件本体和运动控制层面,但随着本体能力的成熟和同质化加剧,价值链的重心逐渐向上游大脑迁移。这个判断在政策端也得到了印证。

2026年政府工作报告和“十五五”规划都明确将具身智能列为国家战略级产业方向。在政策的棋盘上,大脑被赋予了定义产品上限的战略角色。它不再是产业链上游的一个技术模块,已跃升为中游智能底座,占据着整条价值链中最具定价权的生态位。

那么,这个智能底座会朝什么方向进化?

技术架构层面,行业共识正在从单一的VLA模型向信息—物理—认知三域融合演进。世界模型不再是VLA的竞争路线,它是VLA体系中的核心拼图,负责在高维空间里推演;而VLA负责把推演结果转化为具体动作。大脑和小脑有望从各管一摊走向协同进化。

模型能力层面,今天的机器人大脑还停留在被动执行阶段,需要你提供一个指令,它想办法完成;而下一代大脑将具备主动推演和自我进化的能力,会越用越聪明,具有更强的环境适应性和复杂任务处理能力。

部署层面,大脑不能永远长在云端。毕竟,像家庭、商超、医院等场景,没有稳定的网络环境,也不会允许哪怕毫秒级的延迟。而将大算力与大模型部署在端侧,机器人便能逐步摆脱对云端的依赖,在真实世界中完成实时感知与自主决策。

交互层面,未来的机器人或许不再是孤立的个体,它们会逐步适应与人共处的复杂环境,甚至与其他机器人形成群体智能。

因此,从这些维度来说,具身智能大脑的未来,绝不仅仅是更强的算法或更大的模型,它正在长成一个融合认知、推演、控制与学习的全栈式智能中枢。

而当这个中枢真正成熟,机器人将不再是听令行事的工具。它们将成为在开放世界里真正自主理解、自主决策的智能体。

注:文/元君,文章来源:创业最前线(公众号ID:chuangyezuiqianxian),本文为作者独立观点,不代表亿邦动力立场。

文章来源:创业最前线

广告
微信
朋友圈

FAQ回顾

具身智能赛道的融资主要流向哪些领域?

2026年上半年国内具身智能赛道融资近440亿元,超半数资金涌向机器人大脑相关公司;其中VLA模型占大脑赛道融资的42%,其次是世界模型、数据基础设施、端侧芯片,纯本体厂商融资占比仅约12.8%。

当前机器人大脑赛道有哪些主流技术路线?

当前机器人大脑赛道主要有四大主流路线,分别是以VLA与世界模型为核心的模型自研派、侧重端侧部署与量产能力的端侧算力派、兼顾模型与本体的本体大脑一体派,以及从仿生学底层重构架构的类脑架构派。

国内具身智能产业有哪些相关政策支持?

2026年政府工作报告和“十五五”规划已明确将具身智能列为国家战略级产业方向,其中机器人大脑被赋予定义产品上限的战略角色,已跃升为产业链中游智能底座,占据价值链最具定价权的生态位。

未来具身智能市场规模有多大?

赛迪研究院数据显示全球具身智能市场未来五年复合增长率达73%,2030年市场规模有望达2388亿美元;GGII预测2030年国内人形机器人市场规模将接近380亿元,销量将增长至27.12万台。

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