可灵打头阵视频模型IPO潮来了

胡镤心 2026-07-03 11:21

胡镤心 2026/07/03 11:21

邦小白快读

全文速览

这篇文章核心介绍了当前AI视频模型赛道正迎来集体IPO浪潮，可灵AI作为头部玩家打头阵，行业整体呈现出高增长、高竞争、高估值回调的态势，核心干货如下

1. 核心项目进展：可灵AI是全球首个对标Sora的开放影像级视频生成大模型，目前完成30亿美元外部增资，投后估值180亿美元，快手计划12个月内启动其赴港IPO，融资将用于算力建设和核心人才引进。截至2025年底，可灵全球用户突破6000万，年营收超10亿元，累计生成超6亿个视频。

2. 行业整体格局：除可灵外，生数科技、爱诗科技都在筹备IPO，字节跳动的Seedance未分拆独立，目前占据超八成市场份额，港股已经成为AI视频公司IPO的核心主战场。

3. 当前行业风险：AI视频公司普遍保持高增长但同时处于高亏损状态，行业整体高估值正在快速收敛，可灵也遭遇核心创始人离职，竞品强势崛起的挑战。

当前AI视频模型赛道已经进入商业化、资本化爆发阶段，品牌商可从中把握内容营销、产品升级的新机会，核心干货如下

1. 消费趋势变化：AI生成视频已经完成C端用户教育，可灵全球用户破6000万，足以说明普通用户已经普遍接受AI生成内容，品牌做内容营销可借助AI工具大幅降低内容生产门槛，提升内容更新效率。

2. 商业机会：AI视频技术迭代速度快，已经实现原生4K直出、音画同步生成、专业级创作调度等功能，可满足品牌生产广告片、宣传物料、直播素材的各类需求，目前已有超3万家企业客户接入可灵AI服务。

3. 合作风险提示：当前行业技术竞争激烈，核心人才流动频繁，估值也在逐步回调，品牌选择合作方时要优先选择技术稳定、市场份额领先的合作方，避免技术服务中断风险。

AI视频生成赛道正处于商业化增长的红利期，卖家可从中挖掘新的业务机会，同时做好风险规避，核心干货如下

1. 市场机会：当前AI视频生成的C端和B端需求都在高速增长，可灵AI年化收入一年增长4倍，服务企业客户超3万家，足以说明市场需求旺盛，卖家可对接成熟AI模型方，切入AI视频代生成、AI内容定制等To B或To C业务。

2. 当前行业格局：字节Seedance占据超八成市场份额，可灵等新玩家正在突围，多个玩家排队冲刺IPO，整个赛道资金充足，技术迭代快，卖家可选择差异化的细分赛道布局，避开头部垄断领域。

3. 风险提示：当前行业普遍存在高增长高亏损的问题，整体高估值正在快速收敛，核心人才流动频繁，新对手不断出现，竞争烈度高，新手入场要做好成本控制，不要盲目跟风投入重资产。

AI视频模型赛道的快速资本化、商业化，给工厂推进数字化转型、挖掘新商业机会带来诸多启示，核心干货如下

1. 新商业机会：当前品牌、内容端对AI生成视频的需求爆发式增长，工厂可依托自身制造优势，挖掘AI视频赛道的配套商业机会，比如开发适配AI视频创作的智能硬件，或者为中小商家提供AI定制化宣传视频的生产服务，打开新增长曲线。

2. 数字化转型启示：可灵AI从快手内部的一个功能，逐步升级为和电商、商业化并列的一级业务，最终独立分拆冲刺IPO，说明工厂内部孵化的新技术能力，也可以独立出来做商业化，打开新的增长空间。

3. 产业资源对接机会：当前腾讯、阿里、百度等头部产业资本都在布局AI视频赛道，工厂推进数字化转型，可主动对接头部资本和技术方，获得技术和资金支持，加快转型步伐。

当前AI视频赛道处于高速发展阶段，给To B服务商带来很多新的业务机会，核心干货如下

1. 行业发展趋势：AI视频模型技术迭代快速，商业化速度快，可灵AI年化收入一年增长4倍，多个玩家筹备赴港IPO，整个赛道进入资本化爆发期，产业链上下游的配套服务需求旺盛。

2. 客户核心痛点：当前AI视频客户对画质、生成效率的要求不断提升，从基础生成升级到专业级创作，对算力、人才的需求缺口大；同时行业竞争激烈，多数玩家处于亏损状态，面临估值回调、人才流失的多重压力。

3. 解决方案机会：服务商可围绕AI视频赛道，针对性开发算力配套服务、人才寻访服务、企业商业化咨询服务，对接AI模型方和传统企业客户，挖掘To B服务的新增长空间，抓住赛道红利。

AI视频公司集体冲刺IPO，给平台带来新的招商机会，也提出了风险管控的新要求，核心干货如下

1. 招商机会：当前大量AI视频公司有资本化上市需求，除可灵AI外，生数科技、爱诗科技、万兴科技、硅基智能都在筹备上市，港股已经成为AI视频公司IPO的主战场，平台可推出AI赛道专项招商服务，吸引优质AI创业公司入驻。

2. 需要注意的行业问题：当前AI视频公司普遍存在高亏损、高估值的问题，已经上市的智谱AI、MiniMax都出现股价大幅下跌、千亿级别解禁压力的问题，行业整体高估值溢价正在快速收敛，平台需要做好风险防控。

3. 生态建设机会：腾讯、阿里、百度等头部产业资本都在布局AI视频赛道，平台可搭建AI产业对接生态，连接产业资本、AI创业公司和下游客户，提升平台自身的行业影响力和竞争力。

当前AI视频模型赛道迎来集体IPO潮，出现了很多新的产业动向和新问题，具备很高的研究价值，核心干货如下

1. 产业新动向：AI大模型赛道已经从通用大模型竞争，延伸到垂直领域的视频大模型竞争，头部玩家分化出两种发展路径：可灵选择分拆独立融资冲刺IPO，字节选择内部孵化推进商业化，两种路径的发展结果都值得深入研究。

2. 待研究的新问题：当前AI视频公司普遍呈现高增长高亏损的特征，行业出现融资额提升但估值下调的新情况，核心人才流失会直接催生新的竞品，如何给AI公司合理估值、如何做好核心人才留存，都是值得研究的新问题。

3. 资本化路径研究：当前港股成为AI公司上市的核心阵地，智谱AI、MiniMax已经成功上市，为行业树立了定价标杆，可灵AI的IPO结果将成为检验这轮AI资本化浪潮成色的关键，对研究AI产业资本化路径有重要参考价值。

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 centers on the emerging wave of collective initial public offerings (IPOs) in the AI video model sector, with leading player Keling AI leading the charge. The industry overall is characterized by high growth, intense competition, and declining valuations. Key takeaways are as follows:

1. Core project updates: Keling AI is the world’s first open, cinema-quality video generation large model that matches OpenAI’s Sora in positioning. It has completed $3 billion in external fundraising, giving it a post-investment valuation of $18 billion. Kuaishou plans to launch its Hong Kong IPO within 12 months, with proceeds to go toward computing infrastructure expansion and core talent recruitment. By the end of 2025, Keling AI had surpassed 60 million global users, generated over $1.4 billion in annual revenue, and enabled the creation of more than 6 billion videos to date.

2. Industry landscape: Beyond Keling AI, Shengshu Technology and Aishi Technology are also preparing for IPOs. ByteDance’s Seedance remains an integrated business within the company and currently holds over 80% of the market share. The Hong Kong Stock Exchange has emerged as the primary listing venue for AI video companies.

3. Sector risks: AI video companies broadly deliver strong growth while running deep losses, and the industry’s overall high valuations are contracting rapidly. Keling AI itself has faced the dual challenges of a core founder departure and the rise of powerful competitors.

The AI video model track is now entering a phase of explosive commercialization and capitalization, presenting new opportunities for brands in content marketing and product upgrading. Key takeaways are as follows:

1. Shifting consumer trends: AI-generated video has already completed consumer education on the C-end, with Keling AI exceeding 60 million global users confirming that mainstream users now widely accept AI-generated content. Brands can leverage AI tools to drastically lower content production barriers and improve content update efficiency for marketing.

2. Commercial opportunities: AI video technology is iterating rapidly and already supports native 4K output, synchronized audio and video generation, and professional creation scheduling, which can meet brands’ diverse needs for producing ads, promotional materials, and live streaming content. More than 30,000 enterprise clients have already integrated Keling AI’s services.

3. Risk notes for cooperation: The industry currently sees fierce technological competition, frequent core talent turnover, and gradually declining valuations. Brands should prioritize partners with stable technology and leading market share to avoid the risk of interrupted technical services.

The AI video generation sector is in a high-growth commercial dividend window, allowing sellers to tap new business opportunities while managing risks. Key takeaways are as follows:

1. Market opportunities: Both C-end and B-end demand for AI video generation is growing rapidly. Keling AI’s annual revenue has quadrupled in one year and it serves over 30,000 enterprise clients, confirming strong market demand. Sellers can partner with established AI model providers to enter B2B or B2C businesses such as outsourced AI video generation and custom AI content creation.

2. Current industry landscape: ByteDance’s Seedance holds over 80% of the market share, while new players including Keling AI are fighting to break through. Multiple players are in the IPO pipeline, leaving the sector well-capitalized and technology advancing quickly. Sellers should pursue differentiated positioning in niche segments to avoid competing with dominant incumbents.

3. Risk notes: The sector broadly faces the dynamic of high growth paired with high losses, overall valuations are contracting rapidly, core talent turnover is frequent, new competitors constantly emerge, and competition is extremely intense. New entrants should implement strict cost controls and avoid overinvesting in heavy assets blindly.

The rapid capitalization and commercialization of the AI video model sector offers valuable insights for factories pursuing digital transformation and exploring new business opportunities. Key takeaways are as follows:

1. New business opportunities: Demand for AI-generated video from brands and content platforms is growing explosively. Factories can leverage their existing manufacturing advantages to tap supporting business opportunities in the AI video ecosystem, such as developing smart hardware tailored for AI video creation, or offering custom promotional video production services for small and medium-sized merchants, to unlock new growth curves.

2. Insights for digital transformation: Keling AI evolved from an internal feature within Kuaishou into a first-tier business on par with e-commerce and general commercialization, before being spun off independently for an IPO. This shows that new technology capabilities incubated internally at factories can also be spun out for independent commercialization to open new growth space.

3. Industrial resource matching opportunities: Leading industrial investors including Tencent, Alibaba and Baidu are already deploying capital in the AI video sector. Factories advancing digital transformation can proactively connect with leading capital and technology providers to secure technical and financial support and speed up their transformation.

The AI video sector is currently in a phase of rapid growth, creating abundant new business opportunities for B2B service providers. Key takeaways are as follows:

1. Industry development trends: AI video model technology and commercialization are advancing quickly. Keling AI’s annual revenue has quadrupled in one year, multiple players are preparing for Hong Kong IPOs, and the entire sector has entered a period of explosive capitalization, driving strong demand for supporting services across the industrial chain.

2. Core client pain points: AI video clients are increasingly demanding higher image quality and faster generation speed, with requirements evolving from basic generation to professional-level creation, creating large gaps in computing power and talent. At the same time, intense industry competition leaves most players unprofitable, facing multiple pressures including valuations contraction and talent turnover.

3. Solution opportunities: Service providers can develop targeted offerings for the AI video sector, including supporting computing services, talent recruitment, and commercial consulting. By connecting AI model developers with traditional enterprise clients, providers can unlock new growth for their B2B businesses and capture the sector’s dividend.

The wave of simultaneous IPO preparations by AI video companies brings new recruitment opportunities for platforms, while also raising new requirements for risk management. Key takeaways are as follows:

1. Recruitment opportunities: A large number of AI video companies are pursuing public listings. Beyond Keling AI, Shengshu Technology, Aishi Technology, Wondershare, and Silicon Intelligence are all preparing for IPOs, with Hong Kong成为 the primary listing venue. Platforms can launch specialized recruitment programs for the AI sector to attract high-quality AI startups to locate on their platforms.

2. Key industry risks to note: Most AI video companies currently operate with heavy losses and elevated valuations. Already listed players including Zhipu AI and MiniMax have seen steep share price declines and face massive unlock pressure from early investors, and the industry’s overall valuation premium is contracting rapidly. Platforms must implement robust risk prevention measures.

3. Ecosystem building opportunities: Leading industrial investors including Tencent, Alibaba and Baidu are all deploying capital in the AI video sector. Platforms can build an AI industry matching ecosystem that connects industrial capital, AI startups, and downstream clients, to boost the platform’s own industry influence and competitiveness.

The AI video model sector is currently experiencing a wave of collective IPOs, bringing many new industry developments and unresolved questions that offer high research value. Key takeaways are as follows:

1. New industry developments: Competition in the large AI model space has expanded from general-purpose models to vertical video-specific large models. Leading players have diverged into two development paths: Keling AI chose to spin off as an independent entity, raise external capital, and pursue an IPO, while ByteDance chose to keep its Seedance model as an in-house incubated business to drive commercialization. The outcomes of both paths merit in-depth research.

2. New unaddressed research questions: AI video companies broadly feature the dynamic of high growth paired with high losses, and the sector is now seeing higher fundraising volumes paired with lower valuations — a new trend. Core talent departures can also directly spawn new competitors. Key open research questions include how to reasonably value AI companies and how to effectively retain core talent.

3. Research on capitalization paths: The Hong Kong Stock Exchange has become the primary listing venue for AI companies, and already-listed players Zhipu AI and MiniMax have set pricing benchmarks for the sector. The outcome of Keling AI’s IPO will be a key test of this wave of AI capitalization, making it a highly valuable reference for research on AI industry capitalization paths.

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.

【亿邦原创】一个年收入10亿出头、两年亏了24亿的AI公司，估值180亿美元——相当于它母公司市值的75%。这是可灵AI的最新身价。

2026年7月2日晚间，快手公告显示，可灵完成总额上限30亿美元的外部增资，投后估值达180亿美元。本轮融资由CPE源峰、国方创投、BlueFive、腾讯、中关村科学城基金（联合国科投资）、中信证券联合领投，阿里云、百度等产业资本以及华策影视、芒果产业投资人等文娱产业方共同参与。腾讯、阿里、百度同步入局，可灵AI独立商业化进程正式开启。据接近交易的人士透露，快手预计在未来12个月内启动可灵AI赴港上市流程，IPO募集资金将主要用于算力与数据中心建设，以及核心技术人才的引进与留存。

在资本市场，可灵打头阵，生数、爱诗紧随其后，字节的Seedance未分拆但展现出强大的商业化能力——一场AI视频模型的IPO浪潮正在形成。

1、快手公告落定，可灵AI独立上市进入倒计时

可灵AI诞生于快影App里的创作功能，2024年6月正式上线，是全球首个面向公众开放、对标Sora的真实影像级视频生成大模型。

2024年7月启动公测并推出付费会员，同年11月推出独立APP；2025年，陆续发布2.0、2.1、2.5 Turbo、O1及2.6等多个版本；2025年12月发布全球首个大一统多模态视频模型“可灵O1”，实现音画同步生成。

2026年2月，可灵3.0系列模型上线，基于All-in-One理念实现从基础生成到专业级创作调度的系统跃迁。同年4月，可灵AI宣布在视频3.0系列中新增原生4K直出功能，成为业内首个实现该技术的AI视频大模型。

在快手内部，可灵AI的战略地位同样在提升。

2025年4月，快手正式成立可灵AI事业部，升级为与电商、商业化并列的一级业务部门，由高级副总裁盖坤负责，直接向创始人兼CEO程一笑汇报。同年8月，快手发布组织架构调整公告，宣布盖坤兼任可灵AI技术部负责人。

2026年6月，快手社区科学线两位核心技术高管李晗、洪立印完成岗位划转，正式加入可灵AI事业部。

商业化方面，2025年可灵AI全年累计营收约10.4亿元；2026年第一季度收入超6.5亿元，同比增长超300%；截至2026年3月，年化收入运行率（ARR）已逼近5亿美元，较一年前增长4倍。截至2025年底，可灵全球用户突破6000万，累计生成超6亿个视频，服务企业客户超3万家。

然而，就在可灵AI即将迎来资本高光时刻之际，核心人才的流失给盛宴蒙上了阴影。张迪，前快手技术副总裁、可灵项目的核心技术创始人，于2025年8月正式卸任离职。近期，张迪带领的团队推出AI视频生成模型“快乐马”（HappyHorse），在多项评测中已展现出与可灵正面竞争的能力，2026年4月登顶Artificial Analysis视频竞技场排行榜后被阿里认领。2025年12月，快手副总裁、基础大模型及推荐模型负责人周国睿也被曝离职。

2、排队者众：不只可灵AI在冲刺

可灵的分拆与IPO筹备，也赶上了AI公司资本化窗口期。

2026年上半年，智谱AI与MiniMax已相继在港交所成功上市，为整个AI赛道树立了定价标杆。智谱于2026年1月8日以116.2港元/股发行价挂牌，成为全球首家以通用人工智能基座模型为核心业务的上市公司。上市后智谱股价持续攀升，6月22日盘中市值一度突破万亿港元。

MiniMax于1月9日登陆港交所，发行价165港元，上市首日收涨109%，总市值超1000亿港元。两家公司随后于6月被纳入恒生科技指数。

可灵也不是唯一在冲刺资本市场的AI视频玩家。生数科技（Vidu）与爱诗科技（PixVerse）均传出IPO计划。据报道，生数科技已完成股份制改造，B轮融资后估值超20亿美元；爱诗科技正与中金、摩根大通等投行合作筹备上市事宜。有“中国版Adobe”之称的万兴科技在推进“A+H”双重上市，主攻数字人智能体解决方案的硅基智能已向港交所递交招股书。港股正在成为AI视频公司IPO的主战场。

当然，高估值背后也暗藏风险。智谱2025年营收仅7.24亿元、全年净亏损47.18亿元；可灵2024年亏损5亿元、2025年亏损扩大至19亿元。7月2日，智谱大跌近17%，市值跌破8000亿港元；MiniMax大跌近15%，创出3月以来调整新低。两家公司还面临合计近千亿港元市值的巨额解禁压力。AI高估值溢价快速收敛。

从可灵的融资历程中，也可以看到AI视频赛道估值逻辑的根本切换。快手4月最初筹划分拆可灵时，目标估值曾设在200亿美元；最终落定的180亿美元，较初始目标缩水约20亿美元。融资额从20亿美元提高到30亿美元，估值却从200亿美元目标回落到180亿美元——钱更多了，价格反而更谨慎了。

与之相对的是字节Seedance，在20026火山引擎FORCE大会上，火山引擎CEO谭待公开表示，字节跳动并未将视频模型业务分拆独立融资，而是选择在集团内部推进商业化。据媒体报道，Seedance 2.0于2026年2月发布后迅速走红，字节随后开启白名单策略，要求企业客户签订年框以获取稳定API服务。

Seedance目前据称占据市场超八成份额，可灵能否在IPO前缩小这一差距，将直接影响其公开市场定价。

3、敲钟之前，市场等待一个答案

当AI视频公司集体走向IPO，资本市场将以怎样的标尺来衡量它们的价值？是技术领先性，是用户规模，是ARR增速，还是最终的盈利可能？

可以确定的是，可灵在快手体内就是个烧钱的无底洞。2026年快手集团整体资本支出预计达260亿元，较上年增加约110亿元，增量几乎全部投向AI算力。而可灵全年收入仅约10.4亿元，占快手总营收不足1%。这种投入产出反映在资本市场上，2025年年报发布后，快手港股盘中一度跌14%；可灵分拆融资的消息传出后，快手股价一度上涨约11%。

与其让它拖累快手财报、把股价往下拽，不如把它分拆出去，独立融资，证明价值。

可灵能否在2027年初成功登陆港股，将成为检验这波浪潮成色的关键一役。

文章来源：亿邦动力