广告
加载中

原阿里千问负责人创业1个月 估值20亿美元;豆包手机硬件负责人离职 | 独角兽月报

IT桔子 2026-07-01 12:06
IT桔子 2026/07/01 12:06

邦小白快读

EN
全文速览

本文是IT桔子出品的6月全球独角兽企业动态月报,整理了本月独角兽阵营的新增退榜、人事融资战略等核心干货信息,方便快速了解创投圈最新动态。

1. 本月中国共有9家新晋独角兽诞生,集中在具身智能机器人、AI大模型、量子计算等热门赛道,头部项目包括:原阿里千问负责人林俊旸创业1个月估值达20亿美元的卜拉格科技,估值超20亿美元、年增速超3000%的AI内容生成公司Liblib AI,估值100亿元做“Token工厂”的AI基建商硅基流动。

2. 现有独角兽最新动态包括:追觅科技精简业务线优化12%人员,Groq融资6.5亿美元扩张AI推理云业务,前米哈游国际化总裁加入Kimi负责出海,字节拆分AI制药业务、豆包手机硬件核心负责人离职,Meta斥资9亿美元投资印度金融科技独角兽Cred等。

本文整理的最新独角兽动态,能为品牌商把握行业趋势、布局产品研发提供参考干货。

1. 消费趋势层面,AI对C端用户需求的渗透速度远超预期,AI内容创作工具已经获得大规模用户,Liblib AI累计用户超3000万,ARR达3亿美元同比增长超3000%,说明C端AI内容生产需求已经爆发,品牌可以布局AI赋能的内容营销、产品设计等方向,贴合用户新需求。

2. 产品研发层面,当前资本高度看好具身智能、世界模型、AI基建、量子计算等新赛道,大量大厂核心人才流出创业,预示这些赛道即将进入快速增长期,提前布局相关方向的品牌将获得先发优势。

3. 成本控制层面,现在大模型推理已经有标准化的MaaS服务,品牌无需自建算力就能按需调用,大幅降低AI产品研发的门槛和成本。

本文披露的最新创投动态,能为卖家提供政策外的行业风向、增长机会和风险提示干货。

1. 增长机会层面,当前AI内容生产、AI基础设施、具身智能三大细分赛道仍处于高速增长期,头部企业都获得了爆发式增长:硅基流动营收同比增长超10倍,Liblib AI同比增长超3000%,赛道增量空间充足,卖家可以切入相关细分领域找机会。

2. 风险提示层面,AI行业已经从早期的扩张泡沫进入聚焦核心的阶段,头部企业如追觅科技在精简非核心赛道、字节跳动拆分非核心AI业务,卖家要避开同质化的非热门赛道,避免盲目投入。

3. 降本机会层面,现在成熟的MaaS大模型推理服务已经面市,卖家做AI相关产品无需自建算力,按需调用就能降低运营成本,可以借助基础设施服务商的能力快速落地产品。

本文整理的AI赛道创投动态,能为工厂把握产品需求、找商业机会、推进数字化转型提供干货启示。

1. 产品生产需求层面,具身智能机器人是当前资本最热门的赛道之一,本月新晋独角兽中有三家都布局该赛道,涵盖仿生机器人、机器人操作系统等方向,这些项目都需要大量硬件配套支持,传统制造工厂可以针对性调整产能,对接赛道需求,打开新的增长空间。

2. 数字化转型启示层面,AI转型的高成本一直是工厂转型的障碍,现在已经有成熟的“Token工厂”MaaS服务,工厂不需要自建大模型和算力,就能按需调用AI能力,大幅降低数字化、智能化转型的门槛和成本。

3. 商业机会层面,AI产业的快速发展带动了数据中心、AI硬件等相关配套需求,工厂可以依托自身制造优势切入AI硬件配套领域,分享AI产业增长的红利。

本文整理的最新独角兽动态,能为服务商把握行业趋势、挖掘客户痛点、找到新方向提供干货参考。

1. 行业发展趋势层面,当前AI赛道的创投热度依旧居高不下,其中AI内容生产服务、AI基础设施服务、具身智能数据服务三个细分赛道增速最快,资本投入力度大,是未来3-5年的核心增长方向,相关服务商可以提前布局。

2. 客户痛点层面,企业落地AI的核心痛点一直是自建算力成本高、技术门槛高,硅基流动的“Token工厂”MaaS模式刚好解决这个痛点,通过标准化封装大模型推理能力,让企业按需调用,已经验证了模式的可行性,值得服务商参考。

3. 新机会层面,具身智能赛道的快速发展催生了对高质量物理AI数据、AI能力评测基础设施的需求,光轮智能刚完成10亿元战略融资就是明确信号,相关服务商可以布局这个蓝海领域。

本文披露的最新产业动态,能为平台商把握产业需求、调整运营方向、规避风险提供干货内容。

1. 产业对平台的需求层面,越来越多AI企业不愿意投入大量成本自建大模型算力,需要平台提供标准化、按需付费的MaaS大模型推理服务,平台可以新增这块业务,满足客户需求,完善自身生态。

2. 招商运营层面,近期AI多个细分赛道诞生了大量优质高估值初创项目,涵盖AI内容创作、具身智能、量子计算、AI基建等多个方向,平台可以针对性对接这些项目,吸引优质项目入驻,提升平台的整体质量。

3. 风险规避层面,当前AI行业已经进入洗牌分化阶段,头部企业都在精简非核心业务、优化人员,平台要警惕盲目扩张非核心AI赛道的风险,避开估值泡沫过高的同质化项目,聚焦资本和市场认可的核心细分领域,降低运营风险。

本文整理的6月独角兽动态,呈现了当前AI创投领域的最新产业动向,为产业研究提供了很多新鲜干货参考。

1. 产业新动向层面,当前AI行业人才流动加剧,大厂核心技术和管理人才纷纷流出创业,本月就有原阿里通义千问负责人、原字节豆包硬件负责人、原米哈游国际化总裁等核心人才变动,人才流动正在带动AI细分赛道的快速创新,这是值得关注的新特征。

2. 新商业模式层面,AI领域诞生了“Token工厂”这种全新的MaaS商业模式,把大模型推理能力标准化封装后按需收费,既解决了企业客户的痛点,也实现了自身的爆发式增长,一年时间服务1万家企业,营收增长超10倍,该模式值得深入研究。

3. 赛道发展新变化,当前资本投资方向已经从通用大模型研发,逐渐转向AI基础设施、AI落地应用、具身智能、AI+科学计算等方向,创投市场开始分化,优质落地项目依然能获得高估值,产业整体从概念阶段向落地阶段推进。

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 is Itjuzi’s June 2024 monthly report on global unicorn company updates. It compiles core key updates including new unicorn entrants, exit cases, personnel changes, financing and strategic developments to help readers quickly catch up on the latest trends in the venture capital space.

1. This month, 9 new unicorns were born in China, concentrated in hot tracks such as embodied intelligent robots, large AI models, and quantum computing. Standout projects include: Bragi Technology, founded by Junyang Lin, former head of Alibaba’s Tongyi Qianwen, which reached a $2 billion valuation just one month after its launch; Liblib AI, an AI content generation company valued at over $2 billion with annual growth exceeding 3,000%; and SiliconFlow, an AI infrastructure provider known as a "Token Factory" valued at RMB 10 billion.

2. Key updates from existing unicorns include: Dreame Technology streamlined its business lines and cut 12% of its workforce; Groq raised $650 million to expand its AI inference cloud business; the former president of international business at miHoYo joined Kimi to lead its global expansion; ByteDance spun off its AI drug discovery business, and the core head of Doubao’s mobile hardware division stepped down; Meta invested $900 million in Cred, an Indian fintech unicorn.

The latest unicorn updates compiled in this report offer actionable insights for brands to identify industry trends and guide product R&D planning.

1. For consumer trends: AI’s penetration into C-end user demand is growing much faster than expected. AI content creation tools have already gained massive user adoption — Liblib AI boasts over 30 million cumulative users and an ARR of $300 million, with year-over-year growth exceeding 3,000%, confirming that C-end demand for AI-powered content production has already exploded. Brands can seize the opportunity by expanding into AI-enabled content marketing, product design, and other aligned areas to meet new consumer demands.

2. For product R&D: Capital is currently highly bullish on emerging tracks including embodied intelligence, world models, AI infrastructure, and quantum computing, with a large wave of core talent exiting large tech companies to found startups. This signals that these tracks are about to enter a period of rapid growth, and brands that lay out early will gain a first-mover advantage.

3. For cost control: Standardized MaaS (Model-as-a-Service) for large model inference is now widely available. Brands do not need to build their own computing power and can access resources on demand, drastically lowering the barrier and cost of developing AI-powered products.

The latest venture capital updates disclosed in this report offer sellers actionable insights into industry trends, growth opportunities and risk warnings beyond policy information.

1. For growth opportunities: Three core AI segments — AI content production, AI infrastructure, and embodied intelligence — are still in a period of rapid high growth, with leading players already achieving explosive expansion: SiliconFlow has recorded over 10x year-over-year revenue growth, and Liblib AI has grown more than 3,000% year-over-year. These tracks still have ample room for incremental growth, and sellers can explore opportunities in relevant niche segments.

2. For risk warnings: The AI industry has moved beyond the early expansion bubble into a phase of focusing on core businesses. Leading players such as Dreame Technology are streamlining non-core business lines, and ByteDance has spun off its non-core AI business. Sellers should avoid homogeneous competition in non-strategic hot tracks and steer clear of blind investment.

3. For cost reduction: Mature MaaS-based large model inference services are now commercially available. Sellers developing AI-related products do not need to build their own computing power; accessing services on-demand helps cut operating costs, and allows sellers to leverage the capabilities of infrastructure providers to launch products quickly.

The AI venture capital updates compiled in this report offer actionable insights for factories to grasp product demand, identify new business opportunities, and advance digital transformation.

1. For product demand: Embodied intelligent robots are currently one of the most popular tracks for venture capital, with three newly minted unicorns this month focused on the space, covering bionic robots and robot operating systems, among other areas. These projects require extensive hardware supporting, so traditional manufacturing factories can adjust production capacity to match the demand of this track and unlock new growth.

2. For digital transformation: The high cost of AI transformation has long been a barrier for factories. With the emergence of mature "Token Factory" MaaS services, factories no longer need to build their own large models or computing power, and can access AI capabilities on demand. This drastically lowers the barrier and cost of digital and intelligent transformation.

3. For business opportunities: The rapid growth of the AI industry has driven up supporting demand for data centers, AI hardware and other related facilities. Factories can leverage their existing manufacturing strengths to enter the AI hardware supporting space and capture a share of the AI industry’s growth dividend.

The latest unicorn updates compiled in this report offer service providers actionable insights to identify industry trends, pinpoint customer pain points, and explore new strategic directions.

1. For industry trends: The AI sector still maintains extremely high venture capital heat, with three niche segments — AI content production services, AI infrastructure services, and embodied intelligence data services — posting the fastest growth and attracting the largest capital inflows. These segments will be the core growth drivers for the next 3 to 5 years, and relevant service providers should lay out early.

2. For customer pain points: The core pain points for enterprises implementing AI have long been the high cost and technical barrier of building proprietary computing power. SiliconFlow’s "Token Factory" MaaS model directly solves this problem by standardizing and packaging large model inference capabilities for on-demand access, and has already validated the viability of the model, making it a valuable reference for service providers.

3. For new opportunities: The rapid growth of the embodied intelligence track has spurred new demand for high-quality physical AI data and AI capability evaluation infrastructure. The recent RMB 1 billion strategic financing for LightWheel Intelligence is a clear signal, and relevant service providers can enter this blue ocean market.

The latest industry updates disclosed in this report offer platform operators actionable insights to grasp industry demand, adjust operational strategy, and mitigate risks.

1. For platform demand: A growing number of AI companies are unwilling to bear the high cost of building proprietary large model computing infrastructure, and instead require platforms to provide standardized, pay-as-you-go MaaS large model inference services. Platforms can add this business line to meet customer demand and complete their ecosystem.

2. For investment and operations: A large number of high-quality, high-valuation startups have recently emerged across multiple AI niche segments, including AI content creation, embodied intelligence, quantum computing, and AI infrastructure. Platforms can proactively connect with these projects to attract high-quality entrants and elevate the overall quality of the platform.

3. For risk mitigation: The AI industry has now entered a phase of consolidation and differentiation, with leading players streamlining non-core businesses and optimizing headcount. Platforms should guard against the risk of blindly expanding into non-core AI segments, avoid homogeneous projects with excessive valuation bubbles, and focus on core niche segments recognized by capital and the market to reduce operational risk.

The June unicorn updates compiled in this article outline the latest industry developments in AI venture capital, providing fresh, valuable insights for industry research.

1. For new industry dynamics: AI is currently seeing intensified talent mobility, with core technical and management talent from large tech companies leaving to launch startups. This month alone saw high-profile personnel moves including the former head of Alibaba’s Tongyi Qianwen, the former core hardware lead for ByteDance’s Doubao, and the former president of miHoYo’s international business exiting their roles. This talent flow is driving rapid innovation in AI niche segments, making it a notable new trend worth close attention.

2. For new business models: A new MaaS business model called the "Token Factory" has emerged in the AI space. It standardizes and packages large model inference capabilities for on-demand pay-as-you-go access, solving a key pain point for enterprise clients while enabling explosive growth for the provider: it now serves 10,000 enterprises just one year after launch, with revenue growing more than 10x. This model merits in-depth research.

3. For new track development shifts: Capital allocation has gradually shifted from general-purpose large model R&D toward AI infrastructure, AI implementation, embodied intelligence, and AI + scientific computing. The venture capital market is now differentiating: high-quality commercially implemented projects can still command high valuations, and the overall industry is moving from the conceptual phase to the implementation phase.

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.

作者吴梅梅

来源 |IT桔子

每个月,独角兽俱乐部的名单都在悄悄重写。

IT桔子「独角兽月报」专栏,聚焦过去一个月里全球独角兽阵营的关键变动:

谁拿到了那张“十亿美元俱乐部”的入场券?

因IPO上市、估值调整等退榜的又有谁?

以及,现有独角兽们有哪些人事、融资、产品、战略大动作?

我们不追热点,只记录事实。以下是值得关注的本月独角兽企业动态。

据IT桔子数据,今年6月,中国一共有9家新晋独角兽诞生,涉及具身智能机器人、AI大模型、量子计算等细分赛道。

从全球范围独角兽公司新动态来看:Groq融资6.5亿美元,加速扩张AI推理云业务,Meta斥资9亿美元投资印度金融科技公司Cred等值得重点关注。

6月新晋独角兽:

1.千问大模型原负责人离职创业,1个月成独角兽

林俊旸在2026年3月4日通过社交平台X(原Twitter)发文宣布卸任,正式离开阿里巴巴通义千问(Qwen)团队。

作为阿里最年轻的P10技术专家,此前,他一直负责千问技术线,自2022年底起主导大模型研发,包括Qwen3.0及多模态Agent框架的开发,其离职消息引发业内广泛关注。

几个月后,林俊旸的创业消息即流出,并且已锁定首轮融资。

据报道,他创办的AI实验室卜拉格科技于5月成立,近日已经顺利完成首轮数亿美元融资,投后估值约20亿美元,红杉中国、高榕创投两大头部VC各砸1亿美元领投,腾讯跟投2000万美元。

卜拉格科技对应英文Pragmatics,也就是语言学中的语用学,瞄准世界模型。

没有公开产品、没有对外发布技术框架,仅凭创始人履历和方向判断,公司刚创立就拿下百亿级估值,林俊旸的这次创业,直接搅动了当下火热的具身智能、世界模型赛道。

2.Liblib AI成多模态图像/短视频生成独角兽

6月18日,继去年10月完成的B轮融资后,Liblib母公司演语科技宣布完成B+轮融资,总额接近3亿美元,投后估值超过20亿美元。

本轮融资由Granite Asia、腾讯、顺为资本联合领投,HT Investment、时代资本参与,高榕资本、蚂蚁集团、渶策资本、明势创投等老股东跟投。

演语科技形成了三条主要产品线:Liblib AI——AI图片创作社区,累计用户超过3000万;星流Agent——AI设计智能体;LibTV——AI短视频创作平台,于2026年3月上线。

三条产品线围绕专业内容生产逻辑展开,形成从素材、生产到交付的完整AI内容链路,实现了社区、模型和生产工具的深度整合。

截至2026年5月,据称公司ARR已达到3亿美元,同比增长超过3000%,成为全球增长最快的AI应用公司之一。

3.AI基建公司硅基流动成为新晋独角兽

6月16日,AGI基础设施服务商硅基流动宣布完成B轮融资,总金额超过20亿元人民币,据IT桔子测算,其投后估值为100亿元。

本轮融资投资方涵盖携程战投、晶科能源控股、金蝶股份、联通新沃、盛奕资本、壁仞战投、蔚来资本、商汤战投、巨人网络、国泰君安创新投、纪源资本、华控基金、中关村科学城等。

硅基流动的核心商业模式是“Token工厂”,即将大模型推理能力标准化封装为基础设施,通过MaaS(Model as a Service)对外提供服务。企业用户无需自建算力,可按需调用Token,大幅降低AI应用门槛和成本 。

过去一年,硅基流动在企业级市场实现爆发式增长:日均Token调用量达数万亿,服务超过1000万用户和1万家企业客户,营收同比增长超10倍,海外市场单月营收达数百万美元。

除以上重点提及的3家外,6月新晋独角兽公司还包括量子计算技术研发商“本源量子”,仿生机器人研发商“星尘智能”,具身智能机器人大脑/操作系统研发商“无界动力”等。

详情可登录IT桔子独角兽俱乐部页面https://www.itjuzi.com/unicorn追踪。

独角兽公司的新动态:

1.追觅精简业务线,启动优化人员计划

据中国经营报报道,追觅相关业务负责人接受采访时表示,近期,追觅科技完成新一轮核心战略调整,正式锁定智能家庭、户外庭院、智能出行、具身智能四大核心赛道,对旗下业务体系进行系统性梳理与优化。

据悉,此次公司总体裁撤比例约为12%。据估算,目前追觅体系(不含一线工厂员工)总人数约2万人,此次优化涉及人数约2400人。这一比例略高于追觅科技每年年中考核约10%的常态化淘汰比例。

2.Groq融资6.5亿美元,加速扩张AI推理云业务

Groq周一宣布,已成功完成6.5亿美元新一轮增长融资,用于扩展其AI推理云基础设施。本轮融资由Disruptive和Infinitum领投,多位现有投资者也选择跟投。

目前,Groq在北美、欧洲、中东及亚太地区共运营13个数据中心,服务超过500万名开发者以及数千家AI原生企业,每周处理的令牌(token)数量高达数万亿。

3.前米哈游国际化总裁金雯怡已加入Kimi

据报道,金雯怡于今年5月底离职米哈游,在6月初便加入了大模型独角兽月之暗面,其将负责Kimi的出海和全球化,未来将base在海外。

另外,界面报道指出Kimi(月之暗面)已开启新一轮融资计划,投前估值达300亿美元,较此前的180亿美元估值有大幅增长。

4.字节拆分AI制药业务,试水AI4S产业化

近期,字节跳动AI制药业务线已启动拆分与独立融资进程。

据悉,拆分后字节仍将控股新公司,AI制药核心团队、核心算法、技术平台和已有管线资产将整体进入新主体。同时,该业务也将继续从火山引擎获得算力支持。

字节AI制药团队成立于2021年,由刘凯负责。

在字节跳动内部,还有一个Seed AI for Science的团队,专注于科学计算领域的前瞻技术探索,围绕生物领域基础模型、量子化学、分子动力学等方向,用AI推动科学领域的研究范式突破。

值得注意的是,此次调整还伴随着少量人员的离职,比如加入Seed 3年的顾全全,他先后在清华大学完成本硕学业,后在伊利诺伊大学攻读博士,并在UCLA计算机科学系任教、取得副教授荣誉。

5.字节豆包手机硬件负责人林夕离职

据Z Finance报道,多方信源确认,字节跳动AI硬件团队Ocean核心成员、豆包手机硬件产品负责人林夕已于近期离职。这是字节自2024年启动AI手机项目以来,首位出走的核心硬件负责人。

报道提到,林夕此前长期任职于华为终端,是Pura X阔折叠手机硬件产品负责人。

6.具身智能数据服务商光轮智能再获融资

继5月宣布数亿元融资后,近日,光轮智能完成新一轮10亿元战略融资。

本轮投资方包括中关村科学城基金、四川发展科创基金、山东发展科创投等政府基金,以及巨人网络、宇信股份、宝通科技、中科产投、量图智策等产业资本及财务投资机构;老股东建投投资、三七互娱、森马投资等继续跟投。

本轮融资将主要用于持续加大物理AI数据与评测基础设施核心技术研发投入,进一步完善面向机器人学习、能力评测与真实场景落地的产品体系,扩大高质量人类行为数据、仿真合成数据与工业级评测能力建设,并与产业伙伴共同推进开放生态建设。

7.Meta斥资9亿美元投资印度金融科技独角兽Cred

外媒报道,Facebook母公司Meta Platforms斥资9亿美元,投资印度金融科技初创公司Cred。此外,Meta同时任命Cred的创始人Kunal Shah为WhatsApp的新主管。

报道称,该笔投资将使Meta获得了Cred约20%的股份。Cred目前的投后估值为45亿美元。

以上内容依据IT桔子独角兽数据及公开可信资料整理。

榜单更迭是创投市场的常态,但每一次进出背后都有值得关注的信号。

IT桔子独角兽数据库持续更新中https://www.itjuzi.com/unicorn

下个月的独角兽阵营,又会有哪些新面孔?我们拭目以待。

注:文/IT桔子,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。

文章来源:IT桔子

广告
微信
朋友圈

FAQ回顾

2026年6月中国新增了哪些AI赛道的独角兽企业?

2026年6月中国AI赛道新晋独角兽包括:原阿里通义千问负责人林俊旸创办的卜拉格科技,投后估值约20亿美元;多模态AI内容服务商演语科技,投后估值超20亿美元;AGI基础设施服务商硅基流动,投后估值100亿元。

硅基流动的商业模式是什么?

硅基流动核心商业模式为“Token工厂”,将大模型推理能力标准化封装为基础设施,通过MaaS对外提供服务,企业用户无需自建算力即可按需调用Token,可大幅降低AI应用门槛和成本。

AI应用公司演语科技有哪些核心产品线?

演语科技核心有三条产品线:累计用户超3000万的AI图片创作社区Liblib AI、AI设计智能体星流Agent、2026年3月上线的AI短视频创作平台LibTV,形成覆盖素材、生产到交付的完整AI内容链路。

字节跳动近期对AI制药业务做了什么调整?

字节跳动近期已启动AI制药业务线的拆分与独立融资进程,拆分后字节仍控股新公司,AI制药核心团队、核心算法、技术平台及已有管线资产整体进入新主体,后续仍可从火山引擎获得算力支持。

这么好看,分享一下?

朋友圈 分享

APP内打开

+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