• 亿邦公众号
    • APP下载

千百度收购本原智数:一家鞋履上市公司 3亿买下了中国AI数据的「隐形冠军」

作者:Uncle C 文章来源:IPO早知道

2026-06-09 11:31

作者:Uncle C 文章来源:IPO早知道

2026-06-09 11:31

邦小白快读

EN
全文速览

总:本文核心曝光了鞋履上市公司千百度以3亿元对价收购AI数据服务商本原智数90.91%控股权的事件,同时梳理了AI数据赛道的最新发展逻辑,核心干货如下:

1. 核心事件信息:千百度6月公布分两步收购方案,完成后将形成鞋履+人工智能数据双主业,收购后本原智数保持独立品牌与运营,千百度提供现金流与上市平台支持。

2. 行业核心变化:当前AI竞争已经从算力、模型架构比拼,转向高质量真实数据的比拼,全球资本已经开始重估AI数据赛道价值,业界将2026年视为具身智能的数据规模化元年,赛道即将迎来爆发。

3. 隐形玩家特点:本原智数深耕高难度专业数据服务十一年,已经实现稳定盈利,服务国内大半头部大模型、具身智能企业,本次收购的跨界属性反而保障了其数据服务的中立性,形成独特优势。

总:本次收购事件折射出传统品牌转型和AI行业发展的新趋势,对品牌商布局第二增长曲线有诸多参考价值,核心干货如下:

1. 转型路径参考:传统品牌寻求AI转型不需要从零开始搭建团队讲故事,可以通过并购成熟盈利的AI赛道标的落地,本次收购将市场对千百度AI转型的预期落地,是从虚到实转型的典型案例。

2. 差异化优势打造:跨界并购不一定都是劣势,本次收购中千百度不做AI模型的跨界属性,反而变成了新业务的护城河,解决了AI数据服务商最核心的中立性问题,值得品牌跨界布局参考。

3. 产业趋势:当前AI已经从概念落地进入产业爆发期,上游数据需求增长明确,传统品牌只要找对结合点,布局AI相关领域能够拿到新的增长红利。

总:本次事件为AI相关领域的卖家指明了新的增长机会和可借鉴的经营逻辑,核心干货如下:

1. 明确的增长机会:AI行业底层竞争已经发生转移,高质量、高难度的专业AI数据成为核心竞争瓶颈,全球资本已经重估AI数据赛道价值,2026年将迎来具身智能数据规模化元年,赛道需求即将爆发,是明确的增量增长市场。

2. 可借鉴的经营逻辑:本原智数避开低门槛、竞争激烈的通用标注红海,深耕高难度、专业级数据服务赛道,依靠真实订单实现稳定盈利,不跟风烧钱换增长,这种模式在AI创业潮中极为稀缺,值得各类卖家参考。

3. 风险提示与解决方案:AI数据服务商最核心的风险是中立性问题,被头部AI大厂控股容易引发客户流失,背靠无竞争关系的资本母体,是解决该问题的有效路径。

总:本次跨界收购事件,给传统工厂推进数字化转型、挖掘新商业机会带来了清晰启示,核心干货如下:

1. 新商业机会:AI产业爆发带动上游AI数据需求快速增长,尤其是真实物理世界交互数据、垂直领域专业标注数据的缺口极大,传统工厂拥有物理场景资源、垂直领域专业能力,可以接入AI数据交付网络,开辟新的收入来源,打造第二增长曲线。

2. 转型路径启示:传统工厂寻求跨界转型,不一定需要从零开始搭建新业务团队,可以参考千百度的路径,通过并购成熟盈利的新赛道企业,依托原有主业的稳定现金流为新业务赋能,实现平稳落地转型。

3. 组织模式参考:千百度收购后保留本原智数独立运营、独立品牌的模式,既能够享受新赛道的增长红利,又不会冲击原有主业的正常经营,也不会影响新业务的原有竞争力,适合传统工厂跨界转型参考。

总:本次事件清晰展现了AI数据服务行业的发展趋势、核心痛点和可行发展路径,对各类AI相关服务商干货如下:

1. 行业发展趋势:AI产业竞争已经从算力、算法比拼转向高质量数据比拼,AI数据服务已经从过去的成本项重新定义为AI时代的战略资产,行业整体价值正在被全球资本重估,高难度专业数据、真实物理数据需求将持续爆发。

2. 客户核心痛点:当前头部大模型、具身智能企业的核心痛点是难以获取高质量、高安全性、可验证的高难度专业数据,上游优质供给存在较大缺口。

3. 可行发展路径:避开低门槛通用标注的红海市场,深耕高难度垂直专业领域,搭建稳定的专家交付网络与自研交付平台,提前完善数据安全合规资质,就能构建难以复制的护城河;同时保持中立性是核心竞争力,背靠无竞争关系的资本方可长期保持优势。

总:本次事件反映了AI产业链的最新需求,给布局AI相关业务的平台带来多方面启示,核心干货如下:

1. 市场需求与招商方向:当前AI产业链上游,具备规模化交付能力、合规安全、保持中立的高质量AI数据服务商极为稀缺,平台布局AI生态可以将这类标的作为重点招商方向,填补生态缺口。

2. 平台运营管理启示:平台做AI生态,保障服务商的中立性是吸引全行业各类客户的核心,平台自身不直接下场做下游AI模型,反而能吸引不同阵营的客户与服务商,打造更健康开放的AI生态。

3. 机会与风险提示:随着大模型向具身智能、世界模型演进,真实物理世界数据需求爆发,平台可以依托自身资源搭建跨领域专业数据交付网络,抓住赛道红利;同时需要提前完善数据安全合规体系,规避合规风险。

总:本次千百度收购本原智数的事件,展现了AI产业和传统企业转型的诸多新动向,值得研究者关注的干货如下:

1. 产业新动向:AI产业竞争的底层逻辑已经发生转移,从算力、算法之争转向高质量源头数据之争,AI数据服务从成本项重新定义为战略资产,全球资本已经开始重估赛道价值,2026将成为具身智能数据规模化元年,上游源头数据供给的价值开始逐步凸显。

2. 创新商业模式:本次跨界并购创造了AI数据服务商的全新资本模式,传统主业提供稳定现金流,不参与下游AI模型竞争的资本背景,既解决了AI企业长期发展的资金问题,又保留了数据服务商最核心的中立性优势,解决了行业内被大厂控股后客户流失的普遍问题。

3. 产业新问题:过去AI赛道资本更多追捧会讲故事的下游模型企业,闷头做硬供给的上游数据企业长期被低估,这种产业资源分配的变化,也值得深入研究产业发展规律的研究者关注。

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 exposes a core industry event: publicly traded footwear company C.Banner has acquired a 90.91% controlling stake in AI data service provider Benyuan Zhishu for RMB 300 million. It also sorts out the latest development logic of the AI data track. Key takeaways are as follows:

1. Core event details: C.Banner announced a two-step acquisition plan in June. After completion, the company will operate two core businesses: footwear and AI data services. Benyuan Zhishu will retain its independent brand and operations after the acquisition, while C.Banner will provide capital support and access to its public listing platform.

2. Key industry shifts: AI competition has shifted from a race for computing power and model architecture to a competition for high-quality real-world data. Global capital has begun revaluing the AI data sector, and the industry views 2026 as the starting year for large-scale deployment of embodied intelligence data, marking an upcoming explosive growth period for the track.

3. Characteristics of the under-the-radar player: Benyuan Zhishu has focused on high-complexity professional data services for 11 years, has achieved stable profitability, and counts most of China's leading large model and embodied intelligence companies among its clients. The cross-industry nature of this acquisition actually preserves the neutrality of its data services, forming a unique competitive advantage.

This acquisition reflects new trends in both traditional brand transformation and AI industry development, offering valuable insights for brands looking to build a second growth curve. Key takeaways are as follows:

1. A reference transformation path: Traditional brands pursuing AI transformation do not need to build teams from scratch or rely purely on narrative. They can achieve tangible results by acquiring mature, profitable AI-focused targets. This acquisition materializes market expectations for C.Banner's AI transformation, serving as a typical case of moving from abstract concept to tangible execution.

2. Building differentiated competitive advantages: Cross-industry M&A is not inherently a disadvantage. In this case, C.Banner's status as a non-AI model developer has become a moat for the new business, solving the core problem of neutrality faced by AI data service providers. This is a valuable reference for brands expanding into cross-industry AI businesses.

3. Industry trend outlook: AI has moved from conceptual hype to an industrial boom, with clear growing demand for upstream data. Traditional brands can capture new growth dividends from AI-related sectors as long as they find the right point of integration with their existing businesses.

This event points to new growth opportunities and actionable business logic for sellers active in AI-related fields. Key takeaways are as follows:

1. Clear growth opportunity: The underlying competition in the AI industry has shifted, with high-quality, complex professional AI data becoming the core bottleneck. Global capital has already revalued the AI data track, and 2026 will mark the first year of large-scale data deployment for embodied intelligence, bringing imminent demand growth and a clear incremental market.

2. Actionable business strategy: Benyuan Zhishu avoided the red ocean of low-threshold, highly competitive general data annotation and focused on the high-barrier track of complex professional data services. It has achieved stable profitability through real customer orders instead of following the capital-burning growth model. This approach is extremely rare in the AI startup wave and valuable for all types of sellers to learn from.

3. Risk warning and solution: The core risk for AI data service providers is the threat to neutrality: being controlled by leading AI giants often leads to customer churn. Backing by a capital parent with no competing downstream business is an effective solution to this problem.

This cross-industry acquisition offers clear insights for traditional factories pursuing digital transformation and exploring new business opportunities. Key takeaways are as follows:

1. New business opportunity: The AI boom has driven rapid growth in demand for upstream AI data, especially for interaction data from the real physical world and professionally annotated vertical industry data, where gaps in supply are already large. Traditional factories hold physical scenario resources and vertical domain expertise, and can access AI data delivery networks to open up new revenue streams and build a second growth curve.

2. A reference transformation path: Traditional factories pursuing cross-industry transformation do not need to build new business teams from scratch. They can follow C.Banner's example by acquiring mature, profitable companies in emerging tracks, and leverage the stable cash flow from their core original business to empower the new business for a smooth transformation.

3. A reference organizational model: C.Banner retained Benyuan Zhishu's independent operations and independent brand after the acquisition. This model allows incumbents to capture growth from the new track without disrupting their original core business or eroding the competitiveness of the acquired new business, making it well-suited for traditional factories pursuing cross-industry transformation.

This event clearly illustrates the development trends, core pain points and viable growth paths for the AI data service industry, with key takeaways for all AI-related service providers as follows:

1. Industry development trends: AI industry competition has shifted from a race for computing power and algorithms to competition for high-quality data. AI data services, once treated as a cost item, have been redefined as strategic assets for the AI era. The overall value of the industry is being revalued by global capital, and demand for complex professional data and real-world physical data will continue to surge.

2. Core client pain points: The top pain point for today's leading large model and embodied intelligence companies is the lack of access to high-quality, secure, verifiable complex professional data, with a large supply gap for high-quality upstream services.

3. Viable development paths: Players can build an hard-to-replicate competitive moat by avoiding the red ocean of low-threshold general annotation, focusing on complex vertical professional fields, building stable expert delivery networks and proprietary delivery platforms, and securing data security and compliance qualifications in advance. In addition, neutrality is a core competitive advantage, and backing by non-competing capital allows players to maintain this advantage long-term.

This event reflects the latest demand shifts across the AI industry chain, offering multi-faceted insights for platforms expanding AI-related businesses. Key takeaways are as follows:

1. Market demand and investment direction: Upstream in the AI industry chain, high-quality AI data service providers with scalable delivery capacity, compliance and security guarantees, and neutral positioning are extremely scarce. Platforms building AI ecosystems can prioritize such targets for recruitment to fill gaps in their ecosystem offerings.

2. Insights for platform operations: For platforms building AI ecosystems, guaranteeing service provider neutrality is core to attracting clients across the industry. When the platform itself refrains from developing downstream AI models directly, it can more easily attract clients and service providers from all competing camps, and build a healthier, more open AI ecosystem.

3. Opportunity and risk warning: As large models evolve toward embodied intelligence and world models, demand for real-world physical data is surging. Platforms can leverage their existing resources to build cross-industry professional data delivery networks to capture growth from the track boom. Meanwhile, they should improve data security and compliance frameworks in advance to mitigate regulatory risk.

C.Banner's acquisition of Benyuan Zhishu reveals multiple new trends in AI industry development and traditional enterprise transformation, which are worthy of researchers' attention. Key insights are as follows:

1. New industry trends: The underlying competitive logic of the AI industry has shifted, from a battle for computing power and algorithms to competition for high-quality source data. AI data services have been redefined from a cost item to a strategic asset, and global capital has begun revaluing the entire track. 2026 will be the starting year for large-scale embodied intelligence data deployment, and the value of upstream source data supply is gradually coming to the fore.

2. Innovative business model: This cross-industry M&A has created an entirely new capital model for AI data service providers. A traditional core business provides stable cash flow, and the capital background of non-participation in downstream AI model competition solves both the long-term funding problem for AI startups and preserves the core competitive advantage of neutrality for data service providers, addressing the widespread industry problem of customer churn after acquisition by large tech groups.

3. New industry research questions: In the past, AI sector capital mostly favored narrative-driven downstream model companies, while upstream data companies focused on hard supply were long undervalued. The ongoing shift in industrial resource allocation is worthy of in-depth research from scholars studying industrial development patterns.

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.

一家闷头盈利、服务了大半个中国大模型圈的真实数据玩家并入上市公司,第一次走到台前。

据IPO早知道消息,6月5日晚间,千百度(1028.HK)发布公告称,拟通过上海千度收购上海本原智数已发行股本总额约83.33%,总代价为人民币1.5亿元;同时,上海本原智数拟向上海千度发行认购股权,相当于经发行认购股权扩大后已发行股本总额约45.45%,总代价为人民币1.5亿元。

紧随收购事项及认购事项完成后,公司将透过上海千度拥有目标公司已发行股本总额约90.91%。

这意味着,交易完成后,千百度将以3亿对价取得国内领先AI数据服务商本原智数的控股权益,并纳入合并报表,正式形成“鞋履 + 人工智能数据”双主业。

一家鞋履港股上市公司,收购一家AI数据公司,乍看略显“违和”。

但若放进过去几年中国AI产业极速生长的那条暗线里,这桩交易恰恰精准落在了一个正被全球资本重估的赛道上,以及这条赛道中一块长期被低估的硬拼图——当模型与算力日趋同质化,高质量、可验证的数据,正成为通往更高阶智能的关键瓶颈。

对以鞋履零售为主业、近年持续寻找第二增长曲线的千百度而言,这更是一次从“讲故事”到“真转型”的关键落子——市场此前对其AI转型的预期,被这一笔实打实、且已盈利的并购稳稳接住。

一条正在被全球重估的赛道

人工智能的竞争,正在发生一次底层转移。过去十年,胜负主要由算力和模型架构决定;但当公开数据被大模型逐渐消化、当算力与算法日益普及,下一阶段的关键瓶颈已不再是「谁的模型更大」,而是「谁能持续为模型提供新的、更难的、来自真实世界的高质量经验」。

在全球,资本已用真金白银为此投票。2025年6月,Meta以约143亿美元投资数据标注龙头Scale AI,对应估值约290亿美元,一举将「AI数据服务」从过去的「成本项」,重新定义为AI时代的「战略资产」。

逻辑很朴素:算力可以采购,算法可以复现,唯独把人类专家的专业判断、把真实物理世界的交互反馈,规模化且可验证地转化为模型可学习的数据——这件事又难、又慢、难以靠资本速成。它是整条产业链里最难、也最难替代的环节。

业界已将2026年视为具身智能的「数据规模化元年」。当一个赛道开始被全球重新定价,市场终究会回头去寻找那个最扎实、最靠谱的源头供给者。

把「最难的活」做成壁垒

公告称,目标集团(本原智数)为国内领先的人工智能数据服务供应商,亦是国内少数同时具备大模型、世界模型及具身智能完整数据服务能力的供应商之一,处于人工智能产业链上游,核心业务在于将人类的专业与认知经验,转化为人工智能模型可供训练及验证的高质量数据。

回望十一年前,整个行业的注意力几乎都压在算力和模型架构上。而本原智数赌的是另一件事:当算力终将充裕、架构终将收敛,真正决定模型上限的,是数据——不是数据的多少,而是数据的难度、质量与可验证性。

这个判断在当年并不性感。而本原智数偏偏又选了最难走的那条路——不做规模大、门槛低的通用标注,而是扎进价值链里技术门槛最高、最难标准化的高难度、专业级数据服务,专接那些别人不愿做、也做不好的任务:组织数学、代码、法律、医疗等领域的专业人员完成专家级标注,构建衡量模型真实能力的评测数据集,为世界模型重构虚拟环境数据,为具身智能采集真实物理世界的交互数据。

十年下来,本原智数把这条最苦的路,走成了最深的壁垒。如今它服务着国内领先的大模型厂商、一线互联网与内容平台,以及领先的具身智能企业,长期承接他们技术难度最高、要求最严苛的数据任务,并在多个高价值数据品类上成为相关客户的独家供应商。

支撑这一切的,是一张厚重的“交付底盘”:全国十大交付中心、数十万行业专家构成的众包协作网络,以及自研的标注、质检与众包平台、一整套数据安全与合规资质——这正是规模化、高安全地交付高难度数据的核心门槛,也是难以被速成复制的护城河。

更难得的是它的盈利能力。公告显示,2024及2025年,本原智数收入分别约1.465亿元及1.562亿元,税后净利分别为710万元及1,110万元;2026年前五个月,收入增长强劲。

在一级市场动辄“融资数亿、烧钱换增长”的AI创业潮里,一家靠真实订单与经营现金流把规模做起来、还能稳定盈利的数据公司,本身就是极稀缺的标的。

同时站在AI数据爆发的两条主线上

真实世界,才是一切前沿AI数据的源头。

一个常被外界忽略的产业常识是:在具身智能的实际交付中,真实数据在数据采购中的占比高达95%以上;而所有仿真数据,最终都要回到真实世界的数据去校准与验证。因此,当行业越来越聚焦“物理世界的数据为什么这么稀缺”时,做源头真实数据采集与交付的玩家,注定将在未来站稳“基座”的位置。

而本原智数,恰好同时押中了驱动这一轮需求爆发的两条主线:

其一,是模型从“模仿人”走向“自我检验、自我提升”,对高难度、可验证数据的需求激增;其二,是AI走出屏幕、进入物理世界,催生了对真机采集数据的全新需求。

作为数据被用于世界模型训练、并为多家头部「具身大脑」公司提供具身数采的供应商,本原智数是国内少数同时站在这两条主线上的玩家。这意味着,无论未来需求向哪一侧倾斜,它都站在受益的一端。

「违和」之处,恰是护城河

这桩交易最先要过的,是“违和感”这一关。一家AI公司并入一家鞋企,关键或许不在“经营”,而在“结构”。

根据公告,收购完成后,本原智数将继续保持独立品牌与独立运营,并实施数据隔离;千百度则以鞋履主业产生的稳定现金流,为其提供资金与上市平台支撑。这意味着,本原智数无须再靠一轮轮融资续命,能够心无旁骛地做长期、难而正确的事。

更微妙的,是中立性这一层。数据公司最怕“站队”——一旦被某家模型大厂控股,竞争对手往往出于顾虑而转投他处。这样的故事,在行业里并不少见。

而这桩并购的“违和感”,在这里恰恰翻转成了优势。由于千百度本身不做AI模型、不与本原智数的任何客户构成竞争,互为对手的“巨头”们,便仍能放心地把数据交给同一家供应商。一个“看似无关”的资本母体,反而成了守护中立性的最佳人选。

这也是本原智数创始人在公开信中反复强调的一点:中立不是一种短期策略,而是对所有客户的长期承诺。

水面之下的「源头」,开始浮出

随着通用大模型向多模态、世界模型与具身智能持续演进,数据需求迎来爆发,本原智数有望成为这轮上行浪潮中的持续受益者。

把时间拉长,这桩交易真正的看点,或许不在「跨界」,而在它揭开的一个产业事实——在喧嚣的AI浪潮里,最先被资本与媒体追逐的,往往是离模型最近、最会讲故事的那批公司;而像本原智数这样离数据源头最近、闷头把最苦的活做成壁垒、还把生意做成了正向盈利的玩家,反而长期沉在水面之下。

此次并入上市公司、就此“浮出水面”,给了本原智数一个迟来的“扩音器”。当一个赛道开始被全球重新定价,资本与产业终究会回头去寻找那个最扎实、最靠谱的源头供给者。

而这个位置,本原智数已经坐了十年。

注:文/Uncle C,文章来源:IPO早知道(公众号ID:ipozaozhidao),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

乐饮创新获百联挚高资本超亿元战略投资 资金将用于咖啡供应链升级与品牌全球化布局

乐饮创新日前完成亿元股权融资,此次B+轮融资由百联挚高资本独家投资。本轮所募资金将主要用于咖啡供应链的品质升级与品类创新,并加速品牌全球化与管理智能化布局。

咖啡 未来零售

2026-06-09 11:02:26