初创公司尽调
尽调报告 AI video generation / application software late-stage private spinout subsidiary 2026-08-17

Kling AI

高速增长的 AI 视频平台,创作者与生产端牵引真实,但治理、监管和估值风险仍把结论压在「跟踪」

Kling 已成为 AI 视频里最清晰的突破型公司之一;但据称 2026 年 7 月价格已经预支大量未来执行成功,风险调整后的判断仍是观察,而不是买入。

封面要素

发布时间 01
June 2024 [CO003]
总部 02
Beijing, China [CO002]
创作者 03
60M+ [CO023]
生成视频 04
600M+ [CO024]
企业用户 05
30K+ [CO024]
2026 年 Q1 收入 06
650 RMBm+ [CO025]
2026 年 7 月投后估值 08
18000 USDm [CO032, CV001]

公司概况

Kling AI 是 Kuaishou Technology 以北京为运营锚点的 AI 视频业务,2024 年 6 月公开发布,2026 年 7 月完成外部融资后,正准备推进更独立的商业化运营。公司销售面向消费者和专业用户的 AI 创作工具,覆盖文生视频、图像生成、多模态编辑、API 接口、协作功能,以及原生 4K 输出、多镜头叙事等更高端的生产工作流。公开披露显示,这个年轻 AI 视频平台扩张速度异常快:创作者超过 6000 万,生成视频超过 6 亿,合作企业用户超过 30000 家,2026 年 Q1 收入超过人民币 6.5 亿元,2026 年 3 月 ARR 约 $500M。与此同时,公开记录在独立治理、单位经济、留存质量和风险常态化上仍留下主要空白,因此应把 Kling 视为已经具备规模、但披露仍不足的后期私营 AI 公司。

官网
kling.ai
成立时间
2024-06-01
创立地点
Beijing, China
总部
Beijing, China
产品
Kling 销售 AI 短视频和图像生成工具,支持文字、图像、音频、视频作为工作流输入,具备原生多模态生成、原生音频、多镜头叙事、团队协作、移动应用,并越来越多加入原生 4K 输出等面向生产的能力。
客户
全球创作者、代理机构、营销人员、影视团队、企业工作组,以及嵌入或使用 AI 视频工作流的开发者。
商业模式
通过订阅、点数、团队套餐、API 用量,以及更高价值的专业 / 企业工作流实现 B2C/B2B 混合变现。
阶段
late-stage private spinout subsidiary
融资情况
2026 年 7 月公开报道显示,Kling 以约 $18B 投后估值融资约 $2.8B,支持其迈向更独立的商业化运营,同时 Kuaishou 保持多数控制权。
[CO001, CO002, CO003, CO023, CO024, CO025, CO026, CO032]

执行摘要

主要优势

  • 对一个 2024 年 6 月才上线的产品来说,Kling 的规模和商业化信号异常强:60M+ 创作者、600M+ 生成视频、30K+ 企业用户,以及 2026 年 Q1 超过 RMB650M 的收入。
  • 产品似乎已经进入真实工作流,而不只是兴趣实验;公开用例覆盖电影、电视、广告和创作者场景,版本节奏也一路推进到 Kling 3.0 和原生 4K。
  • 2026 年 7 月融资带来强外部背书和资本支持,重要战略与财务投资人都进入了分拆轮。

主要风险

  • 相比当前披露的 $500M ARR 锚点,据称 $18B 投后估值要求很高,也远高于公开软件收入倍数。
  • Kling 在战略上仍依赖 Kuaishou;轮后治理、少数股东保护和完整独立运营准备度,公开证据仍薄。
  • 中国 AI 标识、备案、隐私和应用分发义务,会给面向公众的生成式视频服务带来真实运营和估值风险溢价。
  • credits、退款和支持方面的客户摩擦信号说明,增长质量可能弱于表面用户规模。

未决问题

  • 仍需要独立审计口径财务包,验证毛利率、烧钱速度、现金跑道,以及消费者、企业和 API 收入拆分。
  • 2026 年 7 月轮次之后的股权结构表细节、董事会构成、保留事项和其他少数股东治理保护,公开记录仍未说明。
  • 留存质量披露不足:NRR、GRR、队列留存、退款、拒付和企业扩张之间,公开资料尚未对齐。
  • 算力依赖,以及面对出口管制或基础设施冲击的应急方案,在公开材料里仍不够可见。

目录

Chapter 01

01公司概览

1.1 身份、产品表面和商业模式

Kling AI 更像 Kuaishou 旗下旗舰生成式视频与图像业务,而不是一家拥有长篇公开披露的完全独立公司。官方首页把它定位为下一代 AI 视频和图像生成器;Kuaishou 文件则称其为集团自研多模态视频生成大模型,也是更广泛短视频平台内部的重要商业化引擎。公开产品证据显示,这项业务横跨消费者创作应用、面向创作者的社区和提示词教育界面、专业发布节奏,以及以 Kling 3.0 模型家族为核心的更高端生产工具。这个家族把文字、图像、音频和视频输入并入一个工作流,加入原生音频和多镜头控制,并把生成时长延至 15 秒,同时保持角色和元素一致。变现已经在多个层面可见:消费者会员、点数包、更高分辨率和 4K 加价、团队协作,以及企业或生产端采用。核心承保结论是,Kling 已经不只是挂在 Kuaishou 品牌上的研究演示。它已经是一套商业产品栈,拥有独立包装、创作者教育和工作流专门化;只是消费者、企业、API 和补贴增长渠道之间的具体拆分仍未披露。[CO001, CO003, CO005, CO007, CO008, CO009]

KPI 快照表
指标数值 / 状态日期 / 期间置信度缺口 / 备注
母公司 / 控制权Kuaishou Technology;受控子公司2026-07融资后母公司层面控制权仍为 ~68.33%,并未完全分拆
上线日期June 20242024-06官方上线时间清楚;拆分实体的准确设立沿革不清
最新融资近 $3B / >RMB19B2026-07-02部分报道称承诺金额上限更高,最高达 RMB20.45B
最新估值$18B 投后;$15B 投前2026-07投前 / 投后口径报道一致;优先股堆叠未披露
ARR 里程碑$240M ARR2025-12公司定义的 ARR = 月度经营收入 x 12
收入运行率~$500M ARR2026-03管理层称为约数;未发布经审计的独立损益表
2026 年 Q1 收入>RMB650M2026-Q1披露在母公司业绩中,不是 Kling 单独文件
服务创作者60M+2025-12 to 2026-Q1公司披露;未拆分付费与免费
企业用户 / 客户30,000+2025-12 to 2026-Q1关系类型和合同规模未披露
生成视频600M+2025-12 to 2026-Q1输出数量不等于商业化输出
应用商店信号评分 4.7;28K 条评分2026-08单一 iOS 店面快照,不是全球 MAU 指标
消费者套餐锚点$6.99 Standard / $25.99 Pro / $64.99 Premier / $127.99 Ultra2026-08官方博客价格和应用商店展示,会随渠道和促销状态不同

本表混合了公司事实、管理层披露的规模指标,以及公开显示的消费者定价。收入、创作者和企业数量来自公司披露,而非经审计的 Kling 独立财务报表。

[CO001, CO003, CO013, CO014, CO015, CO016]
FO002: Kling AI 公司快照逻辑

流程图展示 Kuaishou 母平台、Kling 多模态模型、创作者与专业工作流以及商业化触点如何连接。

[CO001, CO007, CO013, CO017, CO027, CO029]
FO003: 快照 KPI

截至 2026-08-17 研究日的关键公开信号:估值、变现、创作者规模、企业触达和应用商店表现。

收入、创作者和企业指标来自管理层披露,而不是经审计的独立报表。应用商店数据是店面快照,不是付费用户质量的直接衡量。

[CO020, CO021, CO023, CO024, CO025, CO026]

1.2 领导层、治理和拆分结构

Kuaishou 母公司的治理能见度明显高于 Kling 自身实体边界。Kuaishou 2025 年年报和英文管理层材料清楚列出联合创始人程一笑担任董事长兼首席执行官,联合创始人宿华担任执行董事;年报还披露了卢蓉 2025 年 4 月获任等独立董事变动。这些披露很关键,因为 2026 年 7 月融资被明确描述为迈向独立商业化运营的一步,而不是已经完成分拆。外部报道还显示,轮后 Kuaishou 将保留约 68.33% 控制权。本文审阅的公开文件集中仍缺少后期投资人通常想看到的独立治理包:完整 Kling 董事会名单、拆分业务的具名独立董事、完整管理团队页面,以及超出母公司控制权标题之外的详细股本结构。这一缺口并不否定业务,但会抬高尽调风险,因为投资人是在用仍主要属于上市母公司的治理证据,为一家按独角兽到十角兽价格计价的公司承保。[CO037, CO038, CO039, CO040, CO041, CO042]

领导层与创始人表
人物角色背景创始人-市场匹配或覆盖关键人依赖
Cheng YixiaoKuaishou 联合创始人、董事长兼 CEO年报和管理层页面披露的母公司创始人兼首席执行官掌控 Kling 的战略资本配置、AI 优先级和母公司层面治理高——公开融资和 AI 战略叙事都经过母公司领导层
Su HuaKuaishou 联合创始人兼执行董事长期母公司创始人,仍列为执行董事Kling 走向分拆时,代表创始人延续性和产品 DNA中——创始人影响力可见,但对当前 Kling 叙事的日常对外主导弱于 Cheng
Lu RongKuaishou 独立非执行董事据年报,2025-04-28 获委任的独立董事在母公司层面提供可见独立监督,而 Kling 独立治理仍未披露低-中——仅限母公司层面治理;没有证据显示她进入公开的 Kling 董事会

本表捕捉 Kuaishou 公开披露中可见的治理边界。已审阅公开来源没有给出完整的 Kling AI 独立管理层名单或董事会页面,因此母公司层面监督是目前披露最好的治理视角。

[CO039, CO040, CO041, CO042, CO051]

1.3 融资、所有权和资本形成

Kling 2026 年 7 月融资,是市场把 AI 视频视为独立资产类别、而非更广义生成式 AI 内部功能的最清晰信号之一。多家独立报道将 7 月 2 日这一轮规模放在略高于 190 亿元人民币,有些来源称最高承诺金额约 204.5 亿元人民币,投后估值约 $18B。财团异常广泛且战略意义突出:Kuaishou 继续控股,Tencent 即便押注自身 AI 生态仍然投资,Alibaba Cloud、Baidu、CITIC Securities、BlueFive、CPE Yuanfeng、Guofang Venture Capital、Zhongguancun Science City Fund 和 CAS Investment 也都出现在交易报道中。这一轮的战略含义至少和融资金额一样重要。Kuaishou 与外部报道都把交易描述为独立商业化运营和最终 IPO 准备,意味着私募市场相信,以接近公开股权的方式分离,可能释放 Kuaishou 上市倍数未反映的价值。限制在于,公开记录仍未披露清算优先权、老股转让、债务、员工股权池细节,或最终按投资人拆分的所有权桥。[CO031, CO032, CO033, CO034, CO035, CO036]

利益相关方 / 投资人地图
利益相关方角色控制权或经济重要性公开证据尽调要求
Kuaishou Technology母公司及控股股东融资后仍保留约 68.33% 控制权MLQ 和 Business20 报道;母公司披露把拆分定位为独立商业运营取得 Kling 层面最终投后股权结构表和保留事项
Tencent战略投资人虽运营竞争性 AI 生态,仍投资约 $200MCNBC、Business20 和 MLQ厘清董事会权利、信息权,以及任何渠道 / 商业协议
Alibaba Cloud战略投资人 / 基础设施盟友释放算力和企业分发相关性信号TechNode、Business20、MLQ判断参投是否附带云资源承诺或优先接入
Baidu战略投资人增加另一家中国主要 AI 生态支持者TechNode、Business20、MLQ厘清战略增值与纯财务参投的边界
CPE Yuanfeng联合领投财务投资人本轮锚定机构资本TechNode、Business20、MLQ审核治理权利和后续跟投意愿
BlueFive Capital联合领投方该轮被突出为 AI 视频领域创纪录规模TechNode、Business20核实持股比例,以及资金是否附带地缘政治或分发条件
CITIC Securities联合领投 / 金融机构为 IPO 路径提前增加公开市场和融资可信度TechNode、Business20、MLQ厘清 IPO 准备角色及任何顾问经济安排
投资方:Zhongguancun Science City Fund / CAS Investment国资背景或政策相关资本强化国内战略重要性和政策支持TechNode、Business20、MLQ了解是否存在政策绩效义务或本地化条件

据报道,2026 年 7 月轮包括数十家投资人;本图聚焦公开报道中最反复被点名的资金提供方,而非列出辛迪加每个参与者。

[CO034, CO035, CO036, CO037, CO038, CO050]

1.4 规模、牵引、里程碑和不利背景

对一个 2024 年 6 月才上线的产品而言,Kling 的规模叙事异常强。Kuaishou 官方发布称,Kling 2025 年 3 月年化收入运行率超过 $100M,2025 年 12 月 ARR 超过 $240M,到 2026 年 3 月接近约 $500M ARR。相同官方材料称,截至 2025 年底至 2026 年初,平台服务全球超过 6000 万创作者,生成超过 6 亿条视频,并与超过 30000 名企业用户或客户建立关系。产品里程碑也推进很快:2025 年 12 月 Omni Launch Week,2026 年 2 月 Kling 3.0,一季度上线团队协作功能,2026 年 4 月发布原生 4K。公开商业证据不限于泛泛的创作者营销。Kuaishou 提到参与历史剧 Swords Into Plowshares 和 Hollywood 剧集 House of David;Kling 自有博客则重点介绍 The RealReal 在 Cannes Lions 获奖的 L’Ultimo Uomo Reale,把它作为情绪连续性和角色稳定性比新奇感更重要的案例。反向压力来自合规和披露风险。中国 2025 年 AI 内容标识规则正面覆盖 AI 生成视频和应用分发界面,Kling 向主流和企业渠道推进时,监管负担也随之变成现实。[CO006, CO018, CO019, CO020, CO021, CO022]

里程碑表
日期事件类型金额 / 状态参与方影响
2024-06Kling AI 上线产品公开上线Kuaishou开启后来拆分的 AI 视频业务商业时间线
2025-03ARR 达到 $100M规模$100M ARR 里程碑Kuaishou / Kling AI显示早期商业化速度异常快
2025-09-01中国 AI 内容标识办法生效监管强制标识和元数据规则CAC、MIIT、MPS、NRTA提高 AI 视频产品和应用分发界面的合规负担
2025-12Omni Launch Week产品发布项:Kling Video O1、Image O1、Video 2.6、Digital Human 2.0Kling AI推动 Kling 从单任务生成走向统一多模态工作流
2025-12月收入超过 $20M;ARR 达到 $240M规模$240M ARRKuaishou / Kling AI证实拆分融资前已形成实质商业化
2026-02-05Kling 3.0 模型系列发布产品发布项:Video 3.0、Video 3.0 Omni、Image 3.0、Image 3.0 OmniKuaishou / Kling AI多模态控制、原生音频和 15 秒生成的重大升级
2026-Q1Kling 收入超过 RMB650M,同比增长 >300%规模>RMB650M;ARR 到 3 月约 USD500MKuaishou / Kling AI把 Kling 定位成母公司第二增长曲线和估值锚
2026-Q1Team Plan 和 Baseball Live 特效被重点展示产品最多支持 15 人协作;42 个国家 App Store 排名 #1Kling AI释放协作推进和海外消费者触达信号
2026-04-23原生 4K 视频发布产品Kling 3.0 系列真正 4K 生成Kling AI / 制作伙伴提升对电影、广告和高端创意工作流的适配度
2026-07-02拆分融资宣布融资近 $3B,投后估值 ~$18BKuaishou、Tencent、Alibaba Cloud、Baidu、CPE Yuanfeng、BlueFive、CITIC、Zhongguancun、CAS 等为独立商业运营提供资金,并设定 IPO 式分拆路径

本时间线强调实质改变产品范围、商业化、监管或所有权的公开里程碑。具体拆分法律步骤、内部重组和非公开客户合同不可见。

[CO003, CO006, CO010, CO021, CO022, CO025]
FO001: Kling AI 里程碑时间线

里程碑时间线,突出 Kling 从 2024 年 6 月发布到 2026 年 7 月剥离融资期间,商业化、产品、监管和融资上的关键拐点。

融资报道对公告金额应读作 >RMB19B,还是上限 RMB20.45B 的承诺略有差异。战略结论不变:2026 年 7 月融资达到十角兽规模,且指向分拆。

[CO003, CO021, CO022, CO025, CO026, CO031]
Chapter 02

02市场分析

2.1 界定 Kling 的真实市场边界

估值 Kling 时,不能没有边界地套用整个生成式 AI 经济。公开产品证据把公司放在几个相邻但并不相同的市场:自助式创作者订阅、移动创意工具、企业 API 推理、广告和电商素材生产,以及更高端的电影或剧集视觉工作流。这些都是真实变现表面,但并不让 Kling 有资格把纯文本 copilots、通用云 GPU 支出,或全球 AI 软件栈的每一美元都算进来。更站得住的框架,是可编程短形态视觉内容生成与编辑,并外溢到一致性、运动质量和参考控制很重要的生产级视听工作流。这个边界很关键,因为 Kling 目前最强的公开证据,不是通用企业 AI 平台故事;而是一个多模态视频创作故事,铺在创作者、营销和工作室用例之上。[CM001, CM002, CM003, CM004, CM005, CM006]

市场定义表
细分 / 类别纳入支出排除支出买方 / 付款方相关性
创作者订阅与移动创作网页 / App 会员、积分、创作者社区使用、图生视频和文生视频生成软件供应商无法捕获的广泛社交媒体广告支出个人创作者 / 自助订阅用户核心,因为 Kling 已通过套餐和应用内购买直接向创作者商业化。
企业 API 视频生成基于用量的推理、预付资源包、协商 API 合同、嵌入式产品工作流通用云算力转售或无关模型托管开发者、产品团队、企业 IT / 平台预算核心,因为 API 访问把 Kling 变成可编程基础设施。
营销与电商素材生产产品演示、短视频广告、品牌视觉、产品叙事工作流不含媒体生成的通用营销技术订阅品牌、代理商、电商、增长预算相关性高,因为官方产品指引明确瞄准产品视频和品牌用例。
电影、电视与高端创意制作分镜可视化、VFX 支持、场景生成、合成镜头、视听创意完整工作室预算、流媒体订阅、摄像硬件工作室、制片人、代理商、制作团队高相关,因为 Kuaishou 提到商业电影和电视落地。
广义生成式 AI 经济多模态 AI 扩张带来的相邻期权价值纯文本 AI 助手、通用企业搜索、GPU 硬件收入、云基础设施N/A仅作有用背景;范围太宽,不能当作 Kling SAM。

边界把 Kling 锁定在视觉生成工作流内,同时排除无关 AI 和基础设施资金池,避免夸大可服务需求。

[CM001, CM002, CM003, CM004, CM005, CM006]
FM001: 市场规模测算视角

从广义 GenAI 需求逐层收窄到 Kling 更有防守性的可服务视觉工作流市场。

底层为定性判断,因为公开证据没有按地域或垂直领域拆出干净的 Kling SAM。

[CM008, CM009, CM010, CM013, CM040]

2.2 市场规模测算视角和自下而上的现实

可用市场研究指向同一方向,但数值并不一致。MarketsandMarkets 和 Statista 等广义生成式 AI 报告给出极大且快速增长的宏观背景;Adwave 引用的较窄 AI 视频估算则小得多,但仍增长很快。这种差异不是需要抹平的噪音;它恰恰说明分析师在测量技术栈的不同层级。对 Kling 来说,最有用的解读是分层:全球 GenAI 需求显然在扩张,AI 视频软件已经成为独立商业品类,Kling 本身也已经达到足够收入规模,证明这个市场并非假设。Kuaishou 披露 Kling 2026 年 Q1 收入超过人民币 6.5 亿元、2026 年 3 月 ARR 约 USD500M 尤其重要,因为即使公开 SAM 和地理拆分仍不可得,它也提供了自下而上的证据,说明有意义的可服务市场已经存在。[CM008, CM009, CM010, CM011, CM012, CM013]

TAM/SAM/SOM 或规模测算视角表
发布方年份地理范围数值CAGR方法置信度局限
MarketsandMarkets2026全球USD 185.45B 生成式 AI 市场36.8% (2026-2033)跨模态、应用和基础设施的广义生成式 AI 市场范围远宽于 Kling 实际可服务市场。
Statista2026全球生成式 AI 市场 24.4%+ CAGR24.4%+对 B2B、B2G 和 B2C 生成式 AI 的自上而下建模市场展望增长率有用;类别范围仍然很宽。
Adwave 引用 Grand View Research2025/2033全球AI 视频生成器市场 2025 年为 USD 788.5M,2033 年达 USD 3.44B20.3%专门 AI 视频生成器工具市场更窄且更相关,但依据第三方摘要,而非分析师原始表格。
DigitalApplied2026全球行业 CAGR 34.2%;2025 年 VC 投资 USD 4.7B34.2%基于 Sora 之后市场框架的行业叙事提供方向性背景,但不是一手市场数据集。
Kuaishou / Kling 自下而上验证Q1 2026全球>RMB650M 季度收入;2026 年 3 月 ARR 约 USD500M2026 年 Q1 收入同比增长 300%+公司披露的当前商业化规模证明真实市场捕获,而不是完整 TAM 估算。

这些是视角,不是可相加组件。它们有意保留彼此矛盾的口径,而不是压成一个虚假精确的 TAM。

[CM008, CM009, CM010, CM011, CM012, CM013]
FM002: 市场估算区间

展示不同市场量级区间,说明单一口径 TAM 不能替代 Kling 可触达市场。

各行混合了不同但标注明确的量级,目的是保留口径差异,而不是暗示它们构成同一个概率分布。

[CM008, CM009, CM010, CM013, CM014, CM015]

2.3 买方、预算所有者和采用路径

Kling 面向多种买方画像;他们共同需要更快的视觉生产,但购买理由不同。个人创作者和半专业消费者看重低摩擦订阅、社区发现和移动端易用性。营销、电商和代理团队更在意产能、产品可视化、品牌一致性,以及短形态素材 ROI。电影、电视和高端广告用户重视角色延续、电影感运动和可控的多镜头排序。开发者和产品团队构成第四类买方,因为 API 接入把视频生成变成产品基础设施,而不是独立工具。这些分段还会以不同方式拆开买方、用户和付款方角色。消费者层级里,用户通常直接付款;企业或工作室部署中,用户可能是创作者或开发者,但付款方是营销、制作、产品或 IT 预算所有者。这种分层意味着 Kling 可以同时通过几个需求漏斗增长,但也意味着某一分段的牵引不应被误读为全分段统治力。[CM019, CM020, CM021, CM022, CM023, CM024]

细分市场 / 买方地图
细分买方用户付款方工作流预算负责人采用触发因素
创作者 / 准专业用户个人创作者同买方月订阅用户或应用内购买用户短内容创意到视频生成与二创个人创作者预算低摩擦、新鲜感、适合社交分发的输出,以及可负担试用成本。
营销 / 电商品牌团队、卖家或代理商设计师、营销人员、运营人员营销或电商预算产品演示、社交广告、商品目录叙事、品牌短视频CMO、增长负责人、电商负责人需要提高内容产出速度并降低制作成本。
电影 / 电视 / 高端创意工作室、制片人或创意总监剪辑师、美术、VFX 团队制作预算分镜可视化、合成镜头、快速场景迭代制片人或工作室负责人需要可控的高质量片段和更快迭代。
开发者 / 产品集成产品或平台团队开发者 / ML 工程师IT、产品或平台预算通过 API 把生成能力嵌入应用或工作流CTO、产品负责人、平台负责人需要可编程的视频生成能力,但不想自建模型。
企业团队协作创意运营经理跨职能内容团队部门预算或创新预算用团队功能和共享资产跑协作流程营销运营 / 数字化转型负责人需要多人协作流程、可复用性和治理。

在消费者档位里,买方、用户和付款方往往合一;到了企业、工作室和 API 驱动部署中,三者明显拆开。

[CM019, CM020, CM021, CM022, CM023, CM024]
FM003: 买方 / 客群图

用矩阵展示 Kling 主要客群的买方、付款方、采用触发点和工作流重点如何不同。

[CM019, CM020, CM021, CM022, CM023, CM024]

2.4 增长驱动、约束和估值相关性

Kling 市场的多头逻辑很直接:AI 辅助视频采用正进入主流,品类工具在一致性和视听完整性上持续提升,成本下降扩大了能定期发布视频的用户集合。Kling 还受益于具备巨大短视频触达能力的母公司生态,以及创作者和企业采用的官方证据。怀疑逻辑同样重要。中国合成媒体规则让标识、审核和溯源变成运营要求,而不是可选的信任功能。跨境 AI 规则为全球扩张增加摩擦。竞争强度非常残酷,Runway、Veo、Pika、Luma、Hailuo 和 Alibaba 支持的 Wan 都在质量、价格、工作流或分发上施压。最重要的是,Sora 2026 年关停说明,仅有需求并不能保证经济性可持续。对估值而言,这意味着相关市场有吸引力但不宽容:Kling 的机会足够大,但利润结构、合规执行和分段留存将决定其中多少能转化为持久企业价值。[CM031, CM032, CM033, CM034, CM035, CM036]

增长驱动因素与约束表
驱动因素 / 约束方向时间影响尽调问题
营销人员多数已采用 AI 辅助视频正向当前品类教育成本下降,软件预算更容易说服Kling 增长中,营销团队与创作者各贡献多少?
从新奇工具转向可控工作流基础设施正向当前至中期一致性、参考素材、编辑和 API 支持做得好的平台更受益用户多频繁重复生产流程,而不是一次性试验?
Kuaishou 分发和生态协同正向当前大型短视频生态可能降低获客和产品学习摩擦有多少流量或转化直接来自 Kuaishou 资产?
品类价格快速下压正向 / 负向当前拉动采用,但可能压住毛利率和差异化按模式、分辨率和客群拆分,贡献毛利是多少?
中国合成内容标识规则负向当前提高合规、内容审核、来源证明和应用分发要求目前上线了哪些水印、元数据和审核系统?
跨境 AI 监管和信任顾虑负向当前至中期让全球企业扩张更慢,合同谈判更重Kling 如何为欧盟和美国客户调整合规姿态?
计算强度和模型经济性负向持续品类需求可能比可持续毛利扩张得更快当前推理成本趋势和 GPU 供应商依赖是什么?
快速迭代的竞争对手负向持续买方可以在 Runway、Veo、Pika、Luma、Wan 等平台之间多栖哪些客户队列表现出真正留存,而不是机会性试用?

同一因素既可能拉动需求,也可能伤到经济性;定价是最清楚的例子。

[CM031, CM032, CM033, CM034, CM035, CM036]
FM004: 采用漏斗或价值链图

AI 视频买方评估 Kling 时,从认知到规模化支出的采用路径。

消费者用户可能很快穿过漏斗,而企业和影视买方可能在工作流和治理阶段停留更久。

[CM029, CM030, CM031, CM032, CM033, CM034]

2.5 图表

Chapter 03

03竞争对手

3.1 Kling 实际竞争的位置

Kling 并不只在一条干净赛道竞争。公开表面显示,它同时处在消费者 AI 视频创作、创作者工作流软件、企业 API 视频生成,以及面向广告和娱乐制作的视听工具中。因此,正确的同业集合会随买方变化。面对创作者和效果营销人员,Kling 与 Pika、Luma 以及其他快周转工具竞争。面对更高端的故事、广告和电影工作流,它会撞上 Runway 和 Veo。面对中国相关或成本敏感买方,它还要面对 Wan、Hailuo、Seedance,以及在一个界面后暴露多个模型的路由层。实际含义是,Kling 受益于品类宽度,但不能假定赢下一个基准或一个价格层级就能锁住整个市场。买方通常在为具体任务做选择:人物真实感、镜头控制、音频、成本、集成或产能。当任务需要电影感短形态输出、可信运动和角色连续性,同时价格低于高端西方工具时,Kling 的公开故事最强。[CP001, CP002, CP003, CP004, CP005, CP006]

竞争对手画像表
竞争对手类别规模 / 融资目标客群差异化限制
Kling AI中国多模态视频平台Kuaishou 支持;>60M 创作者;>30K 企业用户;2026 年 3 月 ARR 约 USD500M创作者、营销人员、影视团队、开发者人体运动、参考一致性、原生音频、多镜头叙事和高性价比短视频输出都较强公开证据更能证明原始能力,护城河证据相对薄;定价和切换风险仍然高。
Runway专业创意工作流平台60M+ 创意人士使用;私人估值据广泛报道已达后期规模创意团队、电影人、开发者、企业用户一体化工作区、开发者平台、角色一致性、编辑深度价格更高;在成本敏感的大批量工作负载中,优势不清楚。
Google Veo模型加生态的既有巨头Google 平台分发和企业触达高端创作者、电影人、企业团队原生音频、真实感、提示词遵循、生态整合不明显针对低成本自助式批量生成优化。
Pika创作者优先的社交视频工具迭代快的消费者 / 创作者品牌创作者、社媒团队、快速原型用户表现力强的特效、低摩擦套餐、快速生成与顶级照片级真实感生产的匹配较弱。
Luma智能体式创意工作流平台面向工作流的创意套件,支持团队协作代理机构、团队、准专业用户、生产用户共享上下文、智能体工作流、生产导出、多资产营销活动公开定价和模型细节比工作流定位本身更薄。
Wan / Hailuo / 路由层中国竞争对手和聚合路由层Alibaba 支持的开源压力,加上其他低摩擦路径成本敏感团队、开发者、中国相关买方价格压力、替代模型访问、开源或路由灵活性可能把底层模型访问商品化,削弱供应商锁定。
现状方案 / 内部自建替代方案现有软件栈和内部团队代理机构、品牌、工作室、产品团队没有新增供应商依赖,审批和流程也熟悉AI 迭代速度更慢,生产成本可能更高。

这些行聚焦最影响决策的替代选项,而不是覆盖市场上每一个视频模型。

[CP001, CP002, CP003, CP004, CP005, CP006]
FP001: 竞争定位图

基于证据展示主要对手在工作流深度和性价比可及性上的位置。

序数评分是分析师基于抓取到的公开证据作出的判断;x=工作流深度 / 企业就绪度,y=性价比可及性,各为 1-10。

[CP005, CP008, CP021, CP022, CP023, CP024]

3.2 按买方任务划分的能力差异

把竞争图谱翻译成购买标准后会更清楚。Runway 强调一体化创意套件和开发者平台,因此对希望在同一环境获得编辑和工作流深度的专业团队特别强。Google Veo 在高端企业和电影制作场景尤其危险,因为它把强真实感、原生音频和更广泛的 Google 生态杠杆结合起来。Pika 更偏创作者原生和表达性,社交与特效取向更轻。Luma 把自身定位在代理式创意工作流、并行协作和生产交付上。Kling 的公开相对优势更具体:强人物运动、多镜头叙事控制、参考一致性、多语言音频,以及可支持更高用量的定价。这足以让它留在第一梯队,但公开来源尚未证明它在每个标准上都有清晰的碾压优势。实践中,这个领域更多按工作负载适配度分化,而不是由一个通用排行榜决定。[CP010, CP011, CP012, CP013, CP014, CP015]

功能 / 能力矩阵
购买标准KlingRunwayVeoPikaLuma
人体动作 / 面部真实感中等中等 / 检索来源未证实
原生音频检索来源显示为部分支持检索来源未证实强调本地化 / 音频工作流
角色 / 参考一致性中等强调工作流连续性
编辑 / 工作流深度初步成形借助 Google 工具达到中等中等
创作者友好定价较弱未证实 / 间接中等
开发者 / API 导向已有检索来源显示有限中等
分发 / 生态杠杆Kuaishou 邻近生态带来强杠杆中等Google 生态带来强杠杆创作者品牌中等团队 / 工作流品牌中等

“未证实”指该能力未在检索到的来源集中得到充分证明,并不表示供应商没有这项能力。

[CP010, CP011, CP012, CP013, CP014, CP015]
FP002: 功能广度 / 能力图

主要竞品类别的能力覆盖和强弱。

强度标签概括已检索证据,不代表基准测试精度。

[CP010, CP011, CP012, CP013, CP014, CP017]

3.3 定价、包装和多栖使用

定价进一步说明市场仍高度碎片化。Kling 消费者套餐显著低于高端企业级创意套件;第三方 API 摘要则把其程序化定价放在预算友好到中档区间,取决于质量模式。Runway 订阅层级明显更高,符合其专业定位。Pika 面向创作者保持低摩擦。Luma 更偏团队和工作流价值,而不只是最低生成成本。Veo 往往通过生态接入竞争,而不是简单按标价正面对比。这些差异创造了一个多栖使用市场。团队可以在一个工具里做原型,在另一个工具里制作主视觉镜头,再通过 Kling 或其他低成本路径跑高量社交或产品内容。聚合层降低工程锁定,让切换更容易。结果是,定价重要,但可持续优势来自可重复的工作流价值、买方能感知的边际质量,以及把重复生产迁到别处会有多痛。[CP021, CP022, CP023, CP024, CP025, CP026]

定价 / 包装对比
公司价格 / 单位 / 合同模式包含能力折扣或未知项影响
Kling 消费端工作室2026 年 8 月博客指南显示:$6.99 Standard、$25.99 Pro、$64.99 Premier、$127.99 Ultra积分、1080p、商业使用、更高月度额度;高档位有优先访问App Store 定价锚因渠道不同而异;实际企业定价未知激进的消费者定价支撑创作者获客和大批量试验。
Kling API第三方摘要显示,按模式和质量不同约 ~$0.084-$0.42 / 秒按用量计费的 API 生成、套餐、批量折扣、定制计划官方价格可能变化,上线前必须刷新可编程访问拓宽可服务场景,但成本对比高度取决于具体任务。
Runway$12-$76 / 月消费者档位;检索来源显示更偏专业定位创意套件、积分、工作流工具API 和企业定价在检索来源中未形成清晰标准更高价格匹配专业工作流定位。
Pika$8 Basic、$35 Pro、$95 Fancy,来自独立对比摘要创作者套餐和风格化视频工具官方定价页在检索来源中信息较少低摩擦创作者竞争让入门价承压。
Luma$29.99 / 月,来自对比摘要;官网强调团队 / 工作流工作流、协作、生产交付、多资产营销活动检索来源中官方定价细节有限更多靠系统级生产率竞争,而不是单纯比最低片段成本。
Veo通常通过 Google 生态产品访问,而不是简单的工作室式标价表通过 Google 入口提供原生音频和高端画质这里直接可比的自助定价仍不够透明对企业用户来说,生态杠杆可能超过标价对比。

市场定价分散在订阅、积分、API 和生态访问模式之间,反过来鼓励多栖使用。

[CP021, CP022, CP023, CP024, CP025, CP026]
FP003: 护城河 / 就绪度 KPI

简要展示 Kling 的强项和竞争暴露点。

[CP031, CP032, CP033, CP034, CP035, CP036]

3.4 护城河耐久性和反向证据

Kling 的护城河论点可信,但有条件。Kuaishou 支持带来资本、分发邻近性和真实商业化证据。官方采用指标显示,产品已经远超演示阶段。公开能力证据也表明,Kling 在人物真实感、运动和短形态商业用例上是严肃竞争者。反向因素同样重要。Runway、Veo 等高端对手在工作流和生态力量上很强;Pika、Luma 等创作者导向对手可以靠速度或集成创意系统获胜;中国同业加聚合层则可能压缩定价并降低切换成本。Sora 停运给出的竞争教训是,品类相关性本身不等于防御力。Kling 很可能属于领先梯队,但证据还没有证明客户不能替换它、绕开它,或强势压价。因此,耐久性取决于 Kling 能否比对手更快把亮眼模型表现转化为重复使用、团队工作流嵌入和伙伴分发杠杆。[CP031, CP032, CP033, CP034, CP035, CP036]

护城河耐久性 / 竞争风险登记表
护城河主张威胁严重性缓解措施 / 尽调问题
Kuaishou 分发协同高端或生态型既有巨头仍能赢下企业需求测算实际获客或留存有多少依赖 Kuaishou 入口。
动作真实感和人物表现竞争对手在真实感和音频上进步很快跟踪人物主体和广告视频工作负载的赢单 / 输单原因。
参考一致性和多镜头控制Runway 和 Veo 也在强推可控性要求提供客户证据,说明 Kling 为什么能赢下重复性多镜头工作。
高性价比短视频输出价格下压可能摧毁利润率或迫使折扣要求按模式和客群拆分贡献毛利。
API 加工作室广度聚合层降低集成切换成本中高量化有多少客户只用 Kling,多少通过路由层使用。
商业牵引规模竞争对手的规模和资本可能追上,或投入更多中高跟踪企业队列留存和扩张,不只看原始客户数。
中国相关生态强度监管和地缘政治可能压低跨境采用中高要求披露地域结构和国际管线转化。

最大风险不是没有需求,而是能力未能转化为持久的工作流锁定。

[CP033, CP034, CP035, CP036, CP037, CP038]

3.5 图表

Chapter 04

04财务

4.1 变现表面清晰,即使收入结构不清楚

Kling 不是还在寻找商业模式的产品。公开来源显示,它拥有分层变现栈:通过网页和应用渠道销售消费者会员,按具体生成功能消耗点数,应用内购买,按用量计费的 API 定价,面向更大程序化用户的预付资源包,以及面向更高用量或企业买方的定制安排。公司还通过团队导向工作流功能变现,而不只是一人创作。这个架构很重要,因为它证明多个收入表面已经存在。未知的是结构。公开披露仍没有告诉投资人,消费者订阅、企业合同、API 用量或特定渠道应用内购买各贡献多少收入。这一区分很关键,因为收入质量、毛利率、支持负担和留存画像会在这些表面之间大幅变化。因此,Kling 的变现机制可见,但投资人仍是在推断每台引擎的权重,而不是看到已披露的收入桥。[CI001, CI002, CI003, CI004, CI005, CI006]

收入来源表
收入流机制单位当前数值 / 状态质量尽调问题
消费者会员按月积分和功能权益收取周期性会员费月度会员官方已记录,且正在销售变现入口可见,但实际 ARPU 和流失未知收入中有多少来自 Standard/Pro/Premier/Ultra 会员?
应用内购买App Store 订阅和积分包按套餐 / 积分包App Store 列表可见可能是高流量漏斗入口,但 App Store 抽成影响未知扣除 App Store 抽成和退款行为后,净收入是多少?
API 使用按用量计费和预付套餐按秒 / 按单位第三方公开汇总;官方价格取决于实时页面有潜力扩展为企业 / 开发者变现面API 作业中生产与测试各占多少,各模式毛利率是多少?
企业 / 定制套餐协商资源包和定制解决方案合同 / 套餐公开可见的是销售动作,不是披露合同金额可能拉高 ACV,但公开透明度最低典型企业 ACV、合同期限和续约率是多少?
团队工作流增购协作套餐和工作流功能席位 / 团队套餐语境Team Plan 显示产品入口已存在如果团队采用,可加深留存,但变现细节缺失已有多少付费团队,团队功能带来多少增量收入?

公开记录能看清变现架构,但看不清这些入口之间的收入结构。

[CI001, CI002, CI003, CI004, CI005, CI006]
FI001: 收入模型桥

用户活动如何转化为 Kling 的变现触点。

流程图展示变现触点的逻辑;它不是披露的会计桥接。

[CI001, CI002, CI003, CI004, CI005, CI006]

4.2 定价记录很丰富,但单位经济并不透明

Kling 发布的定价信号足以建模客户支出行为,但不足以证明盈利能力。官方创作者套餐列出月度点数和功能权益,App Store 展示替代渠道价格,第三方 API 摘要提供按秒计费价格表和套餐测算。付费服务条款还明确,点数会过期、不是现金,并可通过前端价格公告更新,这进一步说明变现紧紧围绕虚拟消耗单位管理。但这些都没有回答真正的承保问题:扣除算力、审核、重试和支持后,Kling 服务一段被接受的视频一分钟成本是多少?WaveSpeed 和 CostBench 都指出,每条可用视频的真实成本取决于被拒输出、审核时间和工作流低效,而不只是标价。这个观察在财务上很重要,因为 AI 视频在点数表上看起来便宜,但如果失败率或返工率很高,进入生产时就会变贵。因此,公开定价是分析收入形态的强起点,不是毛利率尽调的替代品。[CI009, CI010, CI011, CI012, CI013, CI014]

定价 / 变现表
入口公开价格信号计费单位可见性影响
Standard 会员2026 年 8 月博客指南显示,$6.99 / 月,含 660 积分订阅 + 积分官方博客创作者的低摩擦入口。
Pro 会员2026 年 8 月博客指南显示,$25.99 / 月,含 3,000 积分订阅 + 积分官方博客核心中档变现锚点。
Premier 会员2026 年 8 月博客指南显示,$64.99 / 月,含 8,000 积分订阅 + 积分官方博客表示更高支出的创作者 / 准专业用户档位。
Ultra 会员2026 年 8 月博客指南显示,$127.99 / 月,含 26,000 积分订阅 + 积分官方博客支撑重度使用客户,以更低积分成本获得用量。
App Store 渠道定价锚抓取列表显示,月费 $10 Standard、$37 Pro、$92 Premier订阅 / IAP独立平台列表渠道价格不同于网站 / 博客视角。
API 价目表第三方摘要显示约 ~$0.084-$0.420 / 秒按用量计费独立摘要程序化业务经济性取决于任务配置,不是一口价。

Kling 有多套价格口径;不能把任何单一口径当作实际混合费率。

[CI007, CI008, CI009, CI010, CI011, CI012]
单位经济模型表
指标数值 / 空值置信度重要性尽调问题
4K 规划参考30 积分 / 秒说明高端质量会消耗明显更多资源4K 模式相较标准模式的毛利率是多少?
API 标准费率~$0.084 / 秒可作为程序化变现的底价参考套餐折扣和被拒输出之后,实际净价格是多少?
API 高级费率最高质量档位约 ~$0.420 / 秒界定最高质量模式的价格上限哪些工作负载会在规模化后实际支付该费率?
积分有效期付费服务条款下,购买或发放的积分有效期为 2 年延后使用和沉淀收益会影响负债与收入确认判断资产负债表上有多少未使用积分余额?
积分可退款性付费服务条款下,不支持提现或反向兑换影响客户摩擦和沉淀收益经济性各渠道的总退款率和净退款率是多少?
毛利率null算力密集型 AI 视频的核心投资判断变量提供按产品入口和质量模式拆分的毛利。
烧钱速度 / 月度现金消耗null即使融资规模很大,该指标仍决定融资依赖度提供月度烧钱速度和拆分后预计现金跑道。

公开资料把价格表讲得较清楚,但对价格背后的运营经济性披露很少。

[CI015, CI016, CI017, CI018, CI019, CI020]
FI002: 单位经济性桥

从标价到每条可用视频真实成本的定性桥接。

该图为定性分析,因为没有公开来源披露 Kling 毛利率或可接受输出率。

[CI015, CI016, CI017, CI018, CI019, CI020]
FI003: 财务估算区间

公开可支撑的财务锚点和明确缺失字段。

各行是公开锚点,不是完整预测模型。

[CI021, CI022, CI023, CI024, CI025, CI026]

4.3 商业化规模真实,资本充足度看起来很强

Kling 已经从实验走到实质收入规模。Kuaishou 披露 2025 年 12 月月收入超过 USD20M,对应 USD240M ARR;随后又披露 2026 年 Q1 收入超过人民币 6.5 亿元,2026 年 3 月 ARR 约 USD500M。这不是玩具数字。它们说明,即使在 2026 年 7 月拆分融资之前,买方已经在有意义的规模上为 AI 视频付费。融资本身也实质性改变了资本图景。多篇报道将这一轮规模放在约 USD2.8B,投后估值 USD18B,Kuaishou 保持控制。再加上 Kuaishou 截至 2026 年 3 月 31 日拥有人民币 1177 亿元可用资金,母公司生态看起来有能力支持激进的产品和商业化拓展投入。限制在于,这些公开数据点都没有披露 Kling 独立现金、月度消耗或拆分后的资金续航。因此,资本充足度方向上看起来很强,但独立运营阶段的精确流动性仍需要管理层披露。[CI021, CI022, CI023, CI024, CI025, CI026]

资本充足性表
账面现金月度烧钱速度现金跑道(月)计划资金用途下一轮融资触发点债务 / 项目融资义务
Kling 独立账面现金未披露未公开披露未公开披露2026 年 7 月融资后,用于扩大产品、独立运营、商业化,以及可能的算力 / GTM 扩张当增长或算力开支超过变现速度和本轮融资款承受能力时检索资料未披露公开债务或项目融资义务
截至 2026-03-31,母公司 Kuaishou 可动用资金合计 RMB117.7B母公司层面流动性,不等于 Kling 独立现金方向上有支撑,但不等同于 Kling 现金跑道过渡期内,母公司仍可支持平台和 R&D 投入拆分后的治理安排可能逐步削弱直接支持未披露 Kling 专属信贷额度
2026 年 7 月融资约 USD2.8B,投后估值 USD18B新增股权资本如果新股资金大多注入业务,方向上意味着现金跑道较强支撑公司转向独立商业运营公开数据仍无法精确核对新股 / 老股拆分或完整股权结构表优先股堆叠未公开披露

资本充足性方向上较强,但独立现金、烧钱速度和资金用途细节仍属私有信息。

[CI021, CI022, CI023, CI024, CI025, CI026]
FI004: 资本强度 / 现金流图

资本、算力和商业化如何共同塑造 Kling 的财务画像。

公开数据支持现金流图的方向,但不足以支撑精确的独立现金流金额。

[CI027, CI028, CI029, CI030, CI031, CI032]

4.4 真正的承保缺口是收入质量和利润率耐久性

核心财务争论已经不是 Kling 能不能变现。它能。真正的争论是,它正在变成哪一种 AI 视频业务。高留存企业工作流平台,和偏消费者的订阅业务,或一个被竞争模型供应持续压价的 API 产品,理应享受完全不同的倍数。公开反向信号在这里很重要。Trustpilot 评论抱怨点数被浪费、退款、账单摩擦和输出不一致。付费服务条款明确规定,点数不是可退款现金价值,定价或权益结构可以变更。品类先例也重要:Sora 停运说明,吸睛需求可以和破坏性经济性共存。这些并不能证明 Kling 存在 Sora 的问题,但确实证明公司仍需要补上利润率证据。管理层披露分段结构、可接受输出经济性、算力支出、续约行为和现金效率之前,投资人应把 Kling 视为商业上已验证、但财务披露不足的公司。[CI033, CI034, CI035, CI036, CI037, CI038]

公开财务缺口表
未披露私有指标影响具体尽调路径
按消费者 / 应用 / API / 企业划分的收入结构无法判断收入质量、集中度或可比公司口径要求提供 2025 年、2026 年 Q1 和当前运行率的收入桥,按细分、地区和渠道拆分。
按模式和工作流划分的毛利率无法判断收入规模是否具备吸引力的经济性要求提供标准 / pro / 4K / 音频模式下的收入成本和算力成本拆解。
烧钱速度和现金跑道无法判断融资依赖或下一轮融资时点要求提供月度烧钱速度、招聘计划、capex / opex 拆分和融资后现金跑道视图。
留存和净收入留存率(NRR)无法区分重复工作流价值和一次性试用要求按客户细分提供队列留存、合同续约和扩张指标。
积分负债 / 沉淀收益会计可能牵涉收入确认和客户负债要求提供递延收入、积分沉淀收益政策和按账龄列示的未使用积分余额。
按渠道划分的退款 / 支持成本计费摩擦可能侵蚀净收入和品牌信任要求按应用商店、网页和 API 渠道提供退款率、拒付率和支持工单成本。

最大的财务未知数不是 Kling 能否产生收入,而是能否产生持久、高效的收入。

[CI032, CI033, CI034, CI035, CI036, CI037]

4.5 图表

Chapter 05

05产品与技术

5.1 以客户工作流理解产品栈

Kling 最适合被理解为一个多模态创意系统,而不是单一文生视频功能。它的公开产品栈包括用于运动生成的 Video 3.0 和 Video 3.0 Omni,用于静态图像创作和电影感参考工作流的 Image 3.0 和 Image 3.0 Omni,移动端和网页端创作界面,社区发现和二创机制,以及面向可编程使用的 API / 开放平台层。从工作流看,这意味着 Kling 至少服务四类任务:创作者构思、短形态营销生产、电商产品可视化,以及更高端的电影或分镜加速。这个栈还刻意跨模态。官方发布说明把文生视频、图生视频、参考到视频、编辑、多镜头排序、原生音频和元素一致性描述为一个创意引擎的组成部分,而不是彼此独立的单点产品。这让 Kling 比新奇生成器更有商业用途,因为用户可以在一个更广的操作环境里,从提示词走到参考、序列,再到最终输出。[CE001, CE002, CE003, CE004, CE005, CE006]

产品模块 / 资产矩阵
模块 / 资产 / 产品线用户状态 / 成熟度差异化尽调缺口
Kling VIDEO 3.0创作者、营销人员、电影制作人已公开发布,并有较广泛描述15s 生成、原生音频、多镜头、主体一致性没有公开可用性或生产可靠性数据。
Kling VIDEO 3.0 Omni高级创作者、团队、工作室已公开发布,定位为控制力更高的档位视频元素参考、语音控制、多镜头、更丰富的参考工作流没有公开定价或单独采用率拆分。
Kling IMAGE 3.0创作者、营销人员、概念艺术家已公开发布灵活的多参考编辑和更高真实感没有公开吞吐量或使用量披露。
Kling IMAGE 3.0 Omni分镜、预演、设计用户已公开发布2K/4K 直接输出、叙事构图、系列模式、电影化控制除公司摘要外,没有公开基准测试方法细节。
Kling 应用 / 社区入口消费者、创作者、专业消费者通过移动应用、网页和社区分发发现、克隆试用、创作、分享、移动端采用闭环没有公开社区参与度或转化漏斗细节。
Kling API / 开放平台开发者、产品团队、企业公开入口可见程序化集成和按量变现公开文档深度不够,低于完整企业尽调所需。

模块按真实客户任务分组,而不只按模型家族名称分组。

[CE001, CE002, CE003, CE004, CE005, CE006]
工作流 / 使用场景表
用户任务当前工作流公司方案可衡量收益局限
创作者构思提示、生成、混剪、延展应用 / 网页工作室与社区工作流创意到视频的循环更快,输出适配社交平台复杂或长场景仍可能出现一致性失败。
电商产品视频参考图、运动提示、4K 导出图生视频、4K 模式、产品演示引导更低成本、更快生成产品素材品牌审核、重试和打磨仍然重要。
影视 / 分镜预演参考图 / 视频加镜头规划工具:Image 3.0 Omni、Multi-Shot、Video 3.0 Omni电影化镜头规划,预演迭代更快没有公开证据显示它已完全取代传统 VFX 或预演流程。
多语言对话场景角色参考加脚本和音频指令Native Audio 支持多语言和口音把视频和对白生成压到一个工作流里质量和治理仍需人工复核。
嵌入式产品功能第三方应用内的程序化视频生成API / 开放平台访问让 Kling 从工作室 UI 变成基础设施公开文档没有完整披露速率限制或企业控制。

公司没有发布覆盖所有使用场景的标准化 ROI 基准,因此收益只能按方向和工作流判断。

[CE007, CE008, CE012, CE013, CE014, CE021]
FE001: 产品架构图

Kling 对外可见的主要产品层。

这是公开产品层图,不是内部基础设施图。

[CE001, CE002, CE003, CE004, CE005, CE006]

5.2 架构和运营工作流

官方产品材料持续围绕统一多模态架构来描述 Kling 3.0。公开发布说明称,该模型整合文字、图像、音频和视频输入 / 输出,同时在一个训练框架内支持多项任务。实际使用中,工作流从提示词、视觉或视频参考,或二者组合开始;再经过时长、模式、宽高比、音频和镜头结构等生成设置;最终返回可延展、细化或复用的视频或图像输出。Kling 的技术重点是可控性,而不是单纯随机性。多镜头分镜、起止帧控制、元素参考、多图参考和语音绑定角色,都把系统推向生产逻辑。Image 3.0 也有自己的架构叙事,强调视觉链式思考式推理、细粒度感知和叙事审美强化。公开来源没有暴露完整模型图、基础设施栈或服务拓扑,但已经足以得出结论:Kling 被设计成一个理解参考的多模态生成平台,而不是狭窄的单用途模型演示。[CE010, CE011, CE012, CE013, CE014, CE015]

技术 / 运营架构表
层 / 流程 / 组件作用依赖风险
统一多模态模型框架覆盖文本、图像、音频和视频的核心生成引擎模型训练、服务基础设施、参考处理黑箱架构让外部很难审计性能和成本。
参考 / 元素控制锚定主体身份、物体、声音和场景元素用户提供的参考素材及其权利参考素材滥用、IP 问题或一致性不足会削弱输出信任。
多镜头编排把提示结构转成镜头序列和节奏提示清晰度、时长预算、镜头逻辑复杂场景仍可能失败或需要重试。
原生音频生成生成对白、环境声和同步音效语言模型、语音生成、唇形同步对齐音频同步或多语言质量可能随场景复杂度波动。
输出和延展层交付 1080p/4K 短片和更长的延展视频渲染容量、存储、应用 / 网页导出入口更长或更高分辨率输出会增加算力和排队压力。
API 和账户层开放程序化访问、定价和用量跟踪身份认证、配额、账户治理、支持流程公开技术细节和企业控制细节不完整。

本表概括公开运营模型,并不声称还原完整内部系统图。

[CE010, CE011, CE012, CE013, CE014, CE015]
FE002: 客户工作流 / 运营流程

创作者或团队如何完成典型 Kling 工作流。

优化阶段需要单列,因为重试和人工审核仍是实际生产使用的一部分。

[CE007, CE010, CE011, CE012, CE013, CE014]
FE003: 关键依赖图

影响 Kling 运营模型的外部依赖。

依赖项由公开产品和政策界面推断。

[CE015, CE028, CE029, CE030, CE031, CE032]

5.3 部署、集成和产品成熟度

Kling 的公开表面显示,它的交付姿态比许多 AI 视频创业公司更成熟。公司有可见的发布说明、API 页面、应用商店分发、社区层、资源页面和工作流指南,描述的是实际执行,而不只是愿景式营销。2026 年 Q1 披露还显示 Team Plan 最多支持 15 名成员,这很重要,因为许多 AI 创意工具停留在单用户新奇感,而真实企业采用需要协作、交接和受控复用。移动应用可把成片视频延长到最长三分钟,博客指南也描述了围绕 4K、电商和提示词结构的实际生产决策。话虽如此,成熟不等于完全具备企业就绪度。公开文档在可用性、正式 SLA、管理员控制、导出治理、安全认证和支持边界上仍不完整。因此,产品成熟度可见但证据不完整:足以证明严肃部署意图,不足以消除可靠性和企业运营上的尽调问题。[CE020, CE021, CE022, CE023, CE024, CE025]

路线图 / 发布 / 开发阶段表
日期 / 阶段功能 / 里程碑状态含义来源
2024-06 发布背景Kling AI 发布历史锚定整体产品成熟度窗口Kuaishou 新闻材料
2025-12Omni Launch Week,包含 Kling Video O1、Image O1、Video 2.6、Digital Human 2.0已发布显示 3.0 前发布节奏很快Kuaishou ARR 新闻稿
2026-02-05Kling 3.0 模型系列发布已发布叙事控制、音频和一致性的重大升级Kuaishou 3.0 发布
2026-02Team Plan 最多 15 名成员已发布释放协作工作流野心Kuaishou Q1 业绩
2026-04-23原生 4K 视频模型推出已发布把产品进一步推向商业级输出Kling 博客 / 发布
2026-08 全面推出Kling 3.0 全面上线当前显示产品持续成熟,并扩大可用性Kling 发布说明

发布节奏是 Kling 最强的公开可见产品信号之一。

[CE019, CE020, CE021, CE022, CE023, CE024]
FE004: 产品成熟度 / 能力图

按主要产品能力展示方向性的成熟度。

成熟度分数反映证据深度,不代表绝对技术质量。

[CE021, CE022, CE023, CE024, CE025, CE026]

5.4 差异化、信任和控制风险

Kling 最好的公开差异化,在于它把几项有商业意义的能力接进一个栈:可信运动、主体一致性、多语言原生音频、多镜头结构,以及越来越高分辨率的输出。这个组合尤其适合广告、产品演示和电影可视化用例。独立比较来源也强化了这一点:产品在人物表现和短形态电影感输出上尤其强。但信任栈只部分公开。法律条款称输出可能不准确、服务可能不可用,用户内容必须满足广泛权利和合规要求。中国内容标识规则围绕披露增加运营职责。隐私材料描述了对上传和生成数据的广泛收集与处理。结果是,产品故事确有创作深度,但控制风险真实存在:提示词失败、审核开销、参考权利管理和治理义务都仍是运营模型的一部分。因此,从企业尽调视角看,Kling 技术上有差异化,但治理仍未补齐。[CE028, CE029, CE030, CE031, CE032, CE033]

信任 / 质量 / 合规表
控制 / 认证 / 质量指标状态范围缺口
输出标识义务条款和中国规则要求面向公众的生成内容和服务运营实际落地细节未完全公开。
隐私和数据处理披露已有公开政策账户、用户内容、支付和使用数据企业数据治理控制仍有部分未披露。
内容审核和权利规则已有公开条款和政策表述用户上传、生成内容、禁止用途运营审核阈值和执法指标未披露。
准确性免责声明已明确披露所有生成输出确认需要人工复核,也降低任何隐含可靠性保证。
服务可用性免责声明已明确披露平台和应用访问未检索到公开 SLA 或状态历史数据集。
应用商店分发控制可通过 Apple / Google 分发观察消费者和创作者渠道不能证明已具备企业采购就绪度。

本表聚焦可见控制和明确提示,不覆盖未成文的内部防护。

[CE028, CE029, CE030, CE031, CE032, CE033]

5.5 图表

Chapter 06

06客户

6.1 分段图谱和进入表面

Kling 的客户基础应按工作流和购买动作分段,而不是被一个未区分的用户标题概括。最大可见表面是通过网页工作室、iOS 应用和 Android 应用进入的自助式创作者;在这里,买方、用户和付款方往往是同一个购买点数或订阅的个人。第二个分段是专业影视团队,制作人或工作室负责人批准预算,导演、VFX 主管和艺术家在生产中使用工具。第三个分段是代理机构、品牌和电商团队,用 Kling 做短形态活动、产品视频和素材迭代。第四个分段是企业工作组,Team Plan 和 30000 名企业用户披露提供了证据;这里的买方是团队或预算所有者,而不是单个创作者。公开材料还通过 API 和发布说明生态展示了开发者或产品团队表面,但该分段的收入质量远不如电影和创作者层级清晰。结果是一张确有宽度的客户地图,但各分段和渠道之间的证据质量很不均衡。[CU001, CU002, CU003, CU004, CU005, CU006]

客户细分表
细分买方 / 用户 / 付款方使用场景规模 / 证据收入 / 战略价值缺口
自助式创作者 / 专业消费者个人创作者通常同时是买方、用户和付款方文生视频、图生视频、试验、创作者发布60M+ 创作者;28K iOS 评分;10M+ Android 下载和 528K 评论最大漏斗入口,可能贡献订阅 / 积分量的大头没有公开的免费转付费、流失率或单创作者消费数据
影视团队制片人或工作室批准支出;导演、VFX 艺术家和剪辑师使用预演、特效镜头、分镜、电影化场景、虚拟制作项目:House of David、Swords Into Plowshares、Raphael、Born of the Tide、MINIBOTS对高端工作流和品牌可信度最有价值的证明合同规模、续约节奏和活跃工作室账户数未披露
品牌、代理商和电商团队品牌或代理商预算负责人付款;创意团队使用工具营销片、产品视频、产品演示视觉、短视频素材The RealReal 营销活动;官方电商工作流指南;4K 博客中的 Obsidian 引述最可能把输出质量转化为重复营销支出的商业细分公开证据以案例为主,ACV 或采购细节很少
企业工作组 / 团队团队负责人或公司预算负责人付款;多名创作者使用产品协作创作、共享工作流、内部审核和素材管理最多 15 名成员的 Team Plan;声称 30K+ 企业用户可能抬高 ARPU,并降低对个人订阅的依赖没有公开席位数、Team Plan 渗透率或企业留存
开发者 / 产品团队开发者或产品负责人购买;终端用户消费下游输出通过 API 把视频生成嵌入其他产品或工作流可见的 API 和发布说明入口证明开发者漏斗存在如果接入生产系统,按量收入可能更有粘性没有公开 API 客户数、用量或收入拆分

这些细分按购买路径和工作流分组,因为公开证据混合了消费者应用使用、具名生产部署和企业协作宣称。

[CU001, CU002, CU003, CU004, CU005, CU006]
客户获取和渠道依赖表
渠道 / 入口客户群公开证据战略价值风险 / 依赖
Apple App StoreiOS 创作者和专业消费者来自 28K 条评分,4.7/5;应用内套餐和点数可见中国以外有明确获客和变现通道政策、排名和支付通道依赖 Apple
Google PlayAndroid 创作者和准专业用户10M+ 下载;528K 条评论;艺术与设计品类畅销榜排名靠前全球触达规模大,评论可见度高政策、计费和排名依赖 Google
Web 工作室基于浏览器的创作者和专业用户官方发布、发布说明和博客生态非移动工作流可跨平台访问流量和转化指标未公开
Team Plan协作型工作组2026 Q1 官方披露:最多 15 名成员把单人创作接到团队账户未公开使用、扩张或续约数据
电影节 / 制作展示渠道工作室、代理机构和电影人Cannes 小组讨论、Variety 报道、具名项目披露高端工作流可信度,也能支撑企业销售叙事可能放大明星营销节点,而不是稳定复用

核心集中度问题不在某个披露的大客户,而在渠道依赖。

[CU007, CU008, CU009, CU010, CU039]
FU001: 客户旅程图

Kling 的客户旅程从公开发现入口开始;只有个人创作转化为反复的项目或团队使用,才会形成持久价值。

[CU006, CU007, CU024, CU036]

6.2 采用规模和具名客户证据

公开采用证据强到值得认真对待。Kuaishou 称,截至 2026 年 2 月,Kling 服务超过 6000 万创作者;截至 2025 年 12 月,生成视频超过 6 亿条,并与超过 30000 名企业用户合作。消费者规模能见度也具体,而非假设。运行日,美国 App Store 显示 28000 条评分、评分 4.7;Google Play 显示下载量超过 1000 万、评论超过 528000 条。具名生产证据更重要,因为它证明使用超出了业余实验。Wonder Project 公开称 Kling 是 House of David 以及之后 The Old Stories Moses 的核心或基准工具。Variety 另行记录了第一季 72 个 AI 辅助镜头,以及在更广专业工具链中使用 Kling。Timeaxis Studios 称 Kling 被整合进 Swords Into Plowshares 全流程;The RealReal 的 Cannes 获奖作品 L’Ultimo Uomo Reale,以及 Raphael、MINIBOTS 等 Cannes 论坛项目,则显示 Kling 正被用于全球品牌和电影工作流。限制在于,这些证据主要证明生产使用和热情,而不是已披露合同价值或续约行为。[CU011, CU012, CU013, CU014, CU015, CU016]

客户增长 / 采用走势表
指标数值日期来源置信度含义缺失分母
创作者规模全球 60M+ 创作者2026-02-05Kuaishou Kling 3.0 发布稿说明 Kling 已远超小众封闭测试未拆分活跃与流失创作者
生成视频量600M+ 个视频2025-12 / 2026-02 披露组合Kuaishou ARR 公告和 3.0 发布稿支撑真实工作负载,而不只是空注册未按付费、免费或企业使用拆分
企业用户 / 合作伙伴30K+ 企业用户 / 客户2025-12 / 2026-02 披露组合Kuaishou ARR 公告和 3.0 发布稿确认消费者之外已有实质 B2B 渗透未定义用户、客户、试点或付费账户
应用排名突破在 42 个国家和地区登上 App Store 第 12026 Q1Kuaishou 2026 Q1 业绩说明全球消费者发现和传播曾集中爆发没有这些获客高峰队列的后续留存数据
iOS 满意度基础4.7/5,来自 28K 条评分2026-08-17 访问Apple App Store大规模可见评分支撑真实使用评分不等同于付费订阅者或留存创作者
Android 公开足迹10M+ 下载;528K 条评论;艺术与设计畅销榜 #22026-08-17 访问Google Play确认 Android 采用规模可观,且有变现潜力下载和评论看不出付费转化或流失
商业化代理指标Q1 收入超过 RMB650M;2026 年 3 月 ARR 约 USD500M2026-03 / 2026-05 披露Kuaishou 2026 Q1 业绩客户使用已转化为可观变现未按消费者订阅、企业合同或 API 使用拆分

由于 Kling 不发布标准 SaaS 客户队列披露,这张表把采用、应用商店和变现信号放在一起。

[CU011, CU012, CU013, CU014, CU015, CU016]
具名客户证据表
客户客群部署 / 用例正式制作 / 试点结果局限
客户:Wonder Project / House of David影视制作在 AI 辅助镜头流水线中使用 Kling,用于起源段落和更广泛的制作工作正式制作第一季使用 72 个 AI 辅助镜头;后续披露称第二季使用的 AI 镜头数量超过 4 倍,且 Kling 处理了大多数 AI 镜头具体支出、合同条款和总制作预算占比未披露
客户:Timeaxis Studios / Swords Into Plowshares中国历史剧制作从预可视化到最终特效画面集成 Kling正式制作Timeaxis 称,在该流水线中,AI 增强工作流比传统 CG 效率高 3-4x效率表述只针对该项目,不是经审计的客户 ROI 研究
客户:The RealReal / Team One / Lipstick品牌 / 代理机构活动制作用 Kling 在 L'Ultimo Uomo Reale 中保持表演、身份和电影感一致性正式制作该活动获得 Cannes Lions 认可,证明品牌工作流可信度公开记录是创意侧证言,不是披露的合同经济性或复购历史
客户:Mateo AI Studio / Raphael长片制作在一部长篇 AI 电影制作中使用 Kling,目标为院线上映制作进行中说明中国以外也在采用,且使用场景从广告延伸到长片野心仍是制作中的验证,不是已完成的长期商业账户
客户:Evolutionary Films / MINIBOTS动画长片合作为一部即将推出的动画长片建立全球独家技术品牌合作试点到制作意向说明 Kling 在为影视制作支持推动专门合作计划证据仍停在合作阶段;尚无公开成果、续约或收入贡献

这些行比泛泛列出客户名更有力,因为每条都绑定具体工作流或项目;但只有 House of David 和 Swords Into Plowshares 给出了清晰的制作细节证据。

[CU017, CU018, CU019, CU020, CU021, CU022]
FU002: 采用 / 部署漏斗

公开证据说明发现可以推进到生产使用,但无法说明这种使用多常转化为持久经常性收入。

[CU011, CU017, CU021, CU025, CU036]
FU003: 客户验证矩阵

具名生产使用和大规模应用采用的证据最强;续费或合同质量可见度最弱。

[CU013, CU017, CU020, CU023, CU024, CU035]

6.3 留存、满意度和采购摩擦

耐久性是 Kling 公开客户记录中最弱的一环。保留来源没有披露 NRR、GRR、logo 流失、付费转化、平均合同期限,或消费者、团队和企业账号之间的收入拆分。公开记录提供的是积极规模代理指标和可见摩擦的混合体。Apple 和 Google 证明 Kling 拥有大型活跃受众,Google Play 评论显示,至少有部分用户认为图生视频质量足够强,值得认真评估。但 Trustpilot 和 Google Play 也保留了围绕点数烧掉、生成结果消失、账单不透明和支持响应差的持续反向模式。Kling 付费服务条款加深了担忧,因为已激活的付费服务不退款,点数是虚拟物品而非储值,自动续订会持续到明确取消。服务协议还表示,产品按现状和可用状态提供,不保证服务不中断。对企业采购而言,隐私和内容许可条款几乎和原始生成质量一样重要,因为客户需要确信上传材料、生成输出和订阅操作不会制造法律、财务或品牌管理问题。[CU025, CU026, CU027, CU028, CU029, CU030]

留存 / 复用 / 满意度表
指标数值 / 空值客群置信度尽调要求
App Store 满意度信号4.7/5,来自 28K 条评分消费者 / 自助创作者按套餐层级拆分付费订阅者、退款率和队列留存
Google Play 公开使用信号10M+ 下载;528K 条评论;评论文本好坏不一消费者 / 自助创作者提供 Android MAU、付费转化和取消 / 拒付率
Trustpilot 满意度信号1.3/5,反复出现计费、点数和支持投诉消费者 / 自助创作者披露投诉解决指标、支持 SLA 表现和退款结果
企业 NRRnull企业团队 / 更大账户按 Team Plan、企业和 API 客户分组披露 NRR
企业 GRR / 流失null企业团队 / 更大账户提供客户数流失、总留存和扩张对收缩历史
平均合同期限null电影、代理机构和企业账户提供合同期限、试点到生产转化和续约节奏
消费者订阅续费率null付费创作者订阅按月付与年付套餐提供续费、取消、降级和重新激活率

公开证据对初始使用很强,对持久性很弱。空值表示该指标未公开披露,不代表它不重要。

[CU025, CU026, CU027, CU028, CU031, CU032]
FU004: 留存 / 重复使用队列代理

公开披露更多集中在采用和项目验证前端;观察周期越长,公开的持久性证据越薄。

仅作方向性代理。数值衡量每个时间窗口里,公开材料对重复使用持续存在的支撑强弱,不代表公司披露的留存率。

[CU025, CU031, CU035, CU040]

6.4 扩张循环和集中度风险

最可信的客户扩张路径,是从创作者实验进入更重的付费使用,再进入团队协作,最后进入生产或企业工作流。官方材料支持这种递进。应用商店把创作和社区功能卖给个人,付费条款定义会员和点数,Team Plan 为最多 15 名成员加入协作创作,具名影视和商业项目则说明 Kling 可以从单人实验走进工作流嵌入。影视和广告证据还暗示第二条扩张循环:一个成功项目转化为更广泛的工作室或代理关系。但集中度风险仍难承保。没有公开来源披露按头部客户、地域、应用商店或渠道划分的收入集中度。因此,公开证据指向的是渠道集中,而不是经典单一 logo 集中。Kling 高度依赖应用商店做全球消费者发现,依赖 Kuaishou 支持的分发和可信度支撑企业叙事,也依赖少数标杆生产引用证明高端工作流合法性。这足以支持可信的先落地再扩张论点,但不足以有把握地建模客户质量。[CU036, CU037, CU038, CU039, CU040, CU041]

扩张和集中度风险表
扩张驱动集中度风险影响尽调路径
免费或低摩擦创作者入口向点数和订阅扩张高度依赖应用商店发现和计费通道排名、定价或平台政策一变,消费者增长可能快速放缓索取新用户渠道结构、付费转化以及扣除平台费后的净收入
付费创作者向 Team Plan 扩张协作席位相对单人订阅的渗透率不清楚团队账户可能显著抬高 ARPU,但收入占比也可能仍小按套餐索取席位数、付费团队客户和扩张历史
电影项目从单片扩展到多片工作室关系客户价值可能集中在少数明星项目,但使用频次低高端验证有战略价值,但收入可能不稳定索取头部影视账户、每个账户项目数,以及续约或复订数据
品牌和代理机构案例向周期性营销工作扩张活动式预算有周期性,也容易受竞争者影响重复商业项目可能有意义,但公开证据仍是个案索取客户清单、活动复购率,以及按代理机构 / 品牌队列划分的 ACV
借 Kuaishou 背书和产品发布扩张企业 / API 采用可能把自然需求与母公司支持的分发或公关热度混在一起母公司背书有助信任和触达,但可能遮住独立客户集中度按母公司来源渠道、直销和自助 / API 索取收入
全球创作者增长跨地区扩张国家集中度和消费者 / 企业结构未披露表面全球规模仍可能掩盖对少数市场或客户类型的暴露按地域、平台和客群索取收入与活跃用户拆分

扩张逻辑站得住,但公开集中度数据不足以支撑对增长质量的判断。

[CU036, CU037, CU038, CU039, CU040, CU041]
Chapter 07

07风险

7.1 监管和法律风险

Kling 运营在全球最具规定性的 AI 治理环境之一,公开证据显示,合规并不止于模糊原则。中国 2025 年标识办法要求对 AI 生成文字、图像、音频、视频和虚拟场景进行显式与隐式标识;应用分发平台必须审核标识材料;按用户要求提供未标识输出的服务提供者必须保存日志至少六个月。更广泛的中国制度也在通过生成式 AI 备案、深度合成规则、数据安全和隐私法律,以及 2026 年标准制定活动继续扩张。这一点很重要,因为 Kling 是面向公众的生成式视频服务,不是内部工具。它自己的法律文件又增加了更多风险。服务协议称产品按现状和可用状态提供,隐私政策确认对上传和生成内容进行广泛存储与处理,条款还要求明确的会员和点数治理。对投资人而言,法律风险因此不只是一个假设性的版权争论。它是在快速收紧的制度下,内容标识、应用审核、数据处理、用户权利和合同摩擦义务的现实组合。[CR001, CR002, CR003, CR004, CR005, CR006]

监管 / 法律风险登记表
规则 / 案例司法辖区状态可能性严重性缓释措施残余暴露尽调路径
AI 生成内容标识和元数据义务中国自 2025-09-01 起生效在产品和导出流程中内置显式 / 隐式标识索取合规架构、水印和审计证据
应用商店核验 AI 标识材料中国已生效中-高维护随时可备案的文件和应用审核材料包中-高索取中国应用分发合规包和近期审核历史
生成式 AI 备案 / 注册和安全评估负担中国已实施且在扩展中-高保持备案纪律和变更管理控制中-高核验 Kling 专属备案 / 注册编号和节奏
数据安全 / 隐私 / 跨境处理义务中国及国际已实施数据映射、同意控制、传输治理、留存纪律索取 PIPL/DSL 合规备忘录、DPA 和传输控制
服务条款中的合同用户权利 / 商业使用模糊性全球已实施中-高厘清付费与免费商业权利,以及客户合同层级中-高对照企业模板和支持政策审阅消费者条款

各行按严重性排序,依据是它们与 Kling 当前公开产品和分发模式的映射有多直接。

[CR001, CR002, CR003, CR004, CR005, CR006]
FR001: 风险热力图

综合风险最高的是监管 / 法律合规、客户信任摩擦,以及对外部渠道和基础设施的依赖。

[CR001, CR012, CR022, CR025, CR032]

7.2 运营、客户和信任风险

运营风险清晰可见,原因不是公司发布了成熟信任中心,而是 Kling 最强的客户证据来自真实消费者和生产使用。面向客户的条款排除不中断可用性保证,已激活的付费服务不退款,应用商店 / 投诉界面保留了围绕点数消耗、生成结果消失、升级混乱、支持沉默和账单争议的反复不满。这些问题比普通消费者投诉更重要,因为 Kling 正试图从病毒式创作者应用迈向专业生产和企业平台。如果用户感觉失败生成也会消耗稀缺点数,或者支持和开票仍薄弱,从实验到重复付费使用的转化可能卡住。同一问题在更大客户那里会放大。公开来源显示,影视和活动团队在可靠性、权利管理和交付一致性都很重要的工作流中使用 Kling;但关于企业 SLA、事故历史、认证或采购材料的公开证据很薄。结果是一个典型扩张风险:产品热度可见,但保证包装看起来弱于大型、风险敏感买方在把供应商更深嵌入生产系统前通常想要的水平。[CR012, CR013, CR014, CR015, CR016, CR017]

运营 / 质量 / 安全风险登记表
失效模式可能性严重性缓释成熟度残余暴露未解决缺口
生成失败或质量漂移消耗客户点数和信心公开资料看为低-中未公开事故历史或点数补偿政策
计费、退款和支持摩擦拖慢向重复付费使用转化公开资料看为低Trustpilot 和应用评论显示持续不满
缺少公开企业保障材料拖慢采购中-高公开资料看为低未看到公开 SLA、认证或信任中心证据
审核、标识和权利控制跟不上产品宽度中-高公开来源显示义务,但不显示控制有效性
运营依赖应用商店和内容审核流程,带来中断风险中-高中-高无公开备用渠道或直接企业分发细节

主要运营问题不是故障可能发生,而是可见的公开缓释措施比商业野心更薄弱。

[CR012, CR013, CR014, CR015, CR016, CR017]
FR002: 风险传导图

合规、支持和基础设施风险可能很快传导到客户信任、变现质量和估值支撑。

[CR003, CR014, CR026, CR033, CR042]

7.3 依赖和执行风险

Kling 的增长故事也高度依赖对手方和执行选择,而这些并不完全由它掌控。2026 年 7 月融资后,Kuaishou 仍持有约 68% 股权,这意味着即便 Kling 准备更独立地运营,拆分业务在战略和财务上仍与母公司绑定。这可能有帮助,但也会给外部投资人制造治理模糊。应用商店仍是关键的全球发现和计费轨道。House of David、Swords Into Plowshares 和 The RealReal 等标杆证据很有价值,但如果下一波具名企业引用没有出现,它们也会带来叙事集中风险。算力和地缘政治暴露则是另一组依赖。美国出口政策仍在变化:即使 2026 年 1 月 BIS 修订后,允许按个案向获批中国客户发放 H200 许可,先进芯片获取仍由政策调解,而非确定可得。即使 Kling 无需公开披露精确 GPU 组合,这一点仍重要。如果拆分业务的成本结构、产品速度或全球可用性依赖受约束的算力供应链,出口、许可和采购变化就可能迅速传导为运营和战略风险。[CR022, CR023, CR024, CR025, CR026, CR027]

合作伙伴 / 依赖风险登记表
依赖对手方角色集中度失效场景严重性缓释措施残余暴露
母公司所有权和战略控制Kuaishou资本、治理、品牌和运营伞拆分独立进度更慢,或与少数投资者目标冲突明确治理权利和分离计划
移动分发Apple 和 Google应用发现、计费和政策把关排名、政策或审核变化削弱获客或变现分散到 Web 和直连渠道
先进算力供应链GPU / 半导体生态训练与推理产能中高出口或许可政策变化收紧获取,或抬高成本多供应商与前置采购规划
战略投资人生态Tencent、Alibaba、Baidu、CPE、CITIC 等资本加分发关系激励冲突或软性依赖扭曲商业化选择中高明确独立交易的商业条款中高
标杆生产案例Wonder Project、Timeaxis、Team One/Lipstick 等验证高端工作流可信度叙事领先度跑在可复制企业管线前面中高把标杆案例扩展成更广泛的案例库中高

增长越依赖第三方许可、基础设施或声誉背书,集中度越高。

[CR022, CR023, CR024, CR025, CR026, CR027]
人员 / 执行风险台账
角色 / 职能依赖或缺口发生概率影响程度缓释措施尽调路径
剥离后领导层 / 董事会组建独立治理架构尚未完整公开设置独立治理与分拆里程碑索取董事会构成、保留事项和 IPO 准备计划
合规 / 法务运营中国及跨境 AI 义务加码,需要更重的控制体系扩充法务、隐私和内容治理人员索取合规组织架构和外部法律顾问覆盖
客户成功 / 支持运营消费者和专业用户投诉显示,支持能力可能落后于扩张目标中高改进支持响应、开票和退款处理流程索取支持 KPI、人员配置和升级 SLA
企业销售 / 解决方案工程公开证据更像宣传声量,弱于重复采购证明中高中高跑出可被引用的企业销售动作,并建立合同纪律索取按消费者、团队、企业和 API 拆分的管线

执行风险主要不来自纯模型研究,而来自 Kling 要成为耐久独立公司所需的运营工作。

[CR028, CR029, CR030, CR031, CR037]
FR003: 依赖关系图

Kling 的执行要靠母公司、应用商店、监管机构、算力政策,以及少数高含金量验证点。

[CR023, CR024, CR025, CR027, CR029]

7.4 财务 / 模型风险和终止标准

财务模型风险在于,一家增长证据明显的公司,按当前价格仍可能很难干净承保。2026 年 7 月这一轮意味着约 $18B 投后估值,而公开一手来源只披露 ARR 和季度收入快照,没有披露独立利润率、消耗或留存。2026 年 Q1 收入超过人民币 6.5 亿元、ARR 约 $500M 令人印象深刻,但并没有回答业务能否在退款、支持、审核和算力成本同步上升时,仍维持创作者增长。关于海外收入结构和独立计划的二级报道方向上有用,但仍不能替代经审计的分段财务。这意味着正确风险姿态应对价格敏感且可监控。若客户投诉恶化、应用商店分发受损、中国 AI 合规运营负担加重、出口管制或算力获取收紧,或拆分业务仍更像叙事而非运营独立,投资论点会更快破裂。核心问题不是 Kling 是否足够好、能继续增长,而是在治理和耐久性证据仍不完整时,它的当前风险栈是否正足够快下降,以支撑前沿软件估值。[CR032, CR033, CR034, CR035, CR036, CR037]

缓释与叫停标准表
风险可监测触发项阈值 / 事件行动含义
中国合规负担标识、备案或应用审核不合规任何正式执法、暂停或上架 / 更新失败事件暂停投资判断,直到控制缺口修复
客户信任侵蚀投诉强度,以及未解决的计费 / 支持争议公开评价持续恶化,或拒付 / 退款问题上升下调消费者业务耐久性和利润质量
算力 / 出口风险先进芯片获取收紧,或成本飙升实质性新限制、许可被拒或可见的产品容量约束重切增长和利润率假设
剥离执行独立计划停滞,或治理仍不透明融资轮后治理分离没有实质进展施加更高治理折价,并推迟投资
企业转化风险具名生产案例未能转化为更广泛客户标杆案例没有扩展为可复制案例,或披露企业指标下调高端工作流收入预期
估值风险风险栈改善前,增长或 ARR 放缓相对 Q1 / ARR 轨迹显著放缓,且没有更好的留存或治理证据与其激进进场,不如保持观察和价格纪律

叫停标准应能从后续备案、产品更新、平台状态或客户验证中观察到,而不是只靠私下叙事。

[CR032, CR033, CR038, CR039, CR040, CR041]
Chapter 08

08估值

8.1 建议和价格纪律

核心估值问题不是 Kling 产品是否亮眼,而是截至 2026-08-17 的公开证据,能否支撑按 2026 年 7 月传出的估值入场。按这个标准,答案还不到。CNBC、TechNode 等 7 月报道大体指向一轮约 $2.8 billion 融资,投后估值 $18 billion,同时 Kuaishou 仍保留多数控制权。Kuaishou 自身公告证明商业化已经跑起来:Kling 2026 年 Q1 收入超过 RMB650 million;此前在 2025 年 12 月披露 $240 million 收入运行率后,2026 年 3 月 ARR 约达 $500 million。这样的增长足以让投资人继续关注,也足以支撑相对普通软件公司的估值溢价。但同一组证据还没有证明独立利润率能扛住、客户质量指标足够清楚、分部经济经过审计,或治理风险已经拆掉。当价格已经按前沿软件稀缺性定价,信息缺口比产品兴奋度更关键。因此我的建议是观察,而不是买入:这门生意值得密切跟踪,但当前传出的估值,对执行、监管和披露不确定性留下的安全垫太薄。[CV001, CV002, CV003, CV004, CV005, CV006]

建议摘要表
维度评估证据基础决策含义
建议观察增长、融资和产品验证都是真实的,但公开证据还不足以让人有把握支付据报道 $18B 估值。密切跟踪;不要追当前据报道的价格。
置信度公开证据足以框定区间,但独立财务披露不足,难以高置信判断估值是否公允。只有拿到监管文件级或审计级披露后再重估。
风险评级治理依赖、监管负担、客户摩擦信号和不完整的单位经济模型,会形成相关下行风险。做任何新股投资决定前,必须加深尽调。
估值立场偏高据报道估值对应的倍数远高于上市 AI / 软件收入基准,约为已披露 2026 年 3 月 ARR 的 36x。需要更强证据,或更低入场价格。
当前估值下的回报画像安全边际有限上行空间存在,但与剩余证据缺口相比,从当前价格出发的合理结果区间不够有吸引力。宁可等待,不接受被动价格。

本评估对价格敏感,而不只判断公司质量;核心问题是当前据报道的市场价格是否有公开证据支撑。

[CV001, CV003, CV006, CV013, CV028, CV029]
投资逻辑 / 反向逻辑表
论点支撑证据什么会改变判断
Kling 是真实平台,不是科研项目。Kuaishou 披露:2026 年 Q1 收入 >RMB650M、3 月 ARR ~$500M、创作者 60M+、视频 600M+、企业用户 30,000+。审计报表确认耐久的不只是规模,还有增长质量。
影视和高端创意验证支撑稀缺性溢价。House of David、Swords Into Plowshares、围绕 Cannes 的创作者工作流,以及 4K / 3.0 版本发布,显示专业场景已认真采用。披露更多企业 logo 和重复案例后,验证不再这么集中。
据报道的估值已经计入大部分上行。$18B 投后估值明显高于简单的上市软件 / AI 收入倍数,约为已披露 3 月 ARR 的 36x。ARR 爬坡大幅加快,或入场价格实质降低,才会改善交易设置。
风险包装仍薄于估值要求。支持投诉、不可退款点数、母公司依赖,以及独立治理 / 经济模型披露不完整,仍然可见。完整安全 / 合规材料、清晰股权结构、留存指标和更干净的客户质量信号,会降低折价。

正向逻辑强在增长和产品验证;反向逻辑强在价格、治理和证据质量。

[CV004, CV005, CV015, CV018, CV019, CV033]
FV001: 投资建议逻辑

从增长验证和融资信号出发,推导到「观察」建议;披露不足和风险组合限制了结论。

供 IC 使用的概念性综合;箭头表示支撑或折价压力,不是正式因果模型。

[CV001, CV003, CV005, CV019, CV020, CV028]

8.2 可比公司组与倍数现实检验

最干净的估值校验,是把 Kling 隐含倍数拿去和已有市值、收入可见度的上市软件及 AI 相关公司比。CompaniesMarketCap 数据显示,截至 2026 年 8 月,Adobe 市值 / TTM 收入约 4.2x,Duolingo 约 5.7x,GitLab 约 7.2x,C3.ai 约 6.2x。这些可比公司并不完美:Adobe 是有规模、能盈利的创意套件龙头;Duolingo 是消费者订阅平台;GitLab 是开发者工作流公司;C3.ai 是企业 AI 软件纯标的,而不是前沿视频模型公司。即便如此,这个区间仍有用,因为 Kling 传出的 $18 billion 投后估值,约等于 2026 年 3 月 ARR 的 36x,明显高于公开市场为已披露软件收入支付的水平。更快增长、更强战略稀缺性和爆款平台潜力,可以解释一部分溢价。但这么大的溢价,需要治理、单位经济、客户持续性和监管控制都拿出异常干净的证据。公开来源目前还没有给到这种精度,所以正确结论不是 Kling 被炒过头,而是这个倍数已经塞进了大量未来执行成功。[CV011, CV012, CV013, CV014, CV015, CV016]

可比估值表
可比对象指标倍数 / 估值 / 状态相关性局限
Kling AI 据报道 2026 年 7 月融资轮投后估值 / 2026 年 3 月 ARR$18.0B 对 ~$500M ARR(~36.0x)公司自身当前最贴近的价格锚。私募估值;ARR 不等于审计收入;未看到独立申报文件。
Adobe市值 / TTM 收入$104.94B / $25.19B(~4.17x)创意软件龙头,可参照专业创作者工具和变现纪律。规模远大于 Kling,且已上市、盈利能力更强。
Duolingo市值 / TTM 收入$6.21B / $1.09B(~5.70x)消费订阅平台,显示强互动在公开市场能获得什么估值。语言学习模式不同于生成视频的经济模型和监管。
GitLab市值 / TTM 收入$7.19B / $1.00B(~7.19x)工作流和开发者平台可比,参照用量驱动软件变现。不是前沿模型公司,算力和政策风险不同。
C3.ai市值 / TTM 收入$1.54B / $0.25B(~6.16x)纯 AI 软件上市参照,显示市场愿意为 AI 品牌收入支付什么价格。企业 AI 应用软件不同于消费者 + 视频生成业务。

这是方向性估值现实校验,不是公平性意见。关键在于 Kling 相对上市收入基准的溢价幅度。

[CV011, CV012, CV013, CV014, CV016, CV017]
FV002: 估值敏感性

以披露的 2026 年 3 月 ARR ~$500M 为锚,套用不同 ARR 倍数时对应的隐含股权价值。

使用披露的 2026 年 3 月 ARR 锚点,约 $500M。这不是折现现金流模型;只是展示在不同 ARR 倍数假设下,据报道本轮估值要求有多高。

[CV003, CV006, CV013, CV017, CV032]

8.3 情景区间、下行与回报测算

情景分析重要,因为可见事实对增长很强,对财务质量较弱。乐观情景下,Kling 把创作者势能转成可持续的企业和 API 扩张,在大约 12 至 18 个月内把 ARR 推过 $1 billion,压低客户摩擦信号,并证明拆分后能以可信的独立治理和合规体系运转。这样可支撑大约 $18 billion 至 $24 billion 区间;按 2026 年 7 月传出的入场价看,上行只是温和到不错,而不是 VC 式非对称。基准情景更保守:ARR 继续增长,但收入结构仍混在消费者积分、订阅和仍在成熟的企业用量之间;治理和风险包装只部分改善;市场仍给溢价,但不再给英雄式倍数。这条路径下,约 $10 billion 至 $15 billion 更站得住。悲观情景并不牵强。如果监管加重、应用商店或客服摩擦恶化、出口管制或算力约束收紧,或风险正常化前增长已经降速,估值可能压缩到 $6 billion 至 $10 billion。也就是说,当前传出估值只有在多个有利假设同时兑现时才成立;只要少数风险叠加,下行就会很快打开。[CV022, CV023, CV024, CV025, CV026, CV027]

乐观 / 基准 / 悲观情景表
情景假设估值 / 回报逻辑关键风险概率信号
乐观ARR 在 ~12-18 个月内超过 $1B;企业 / API 占比加深;治理和合规材料显著改善;客户摩擦指标稳定。$18B-$24B;相对据报道 $18B 融资轮,约为持平到 ~1.3x 上行,因此即使乐观情景相对当前价格也不算大。必须跑通企业质量,而不只是增长规模。可能发生,但需要多个有利里程碑同时落地。
基准ARR 继续增长至约 $650M-$800M 区间;创作者与企业收入组合仍然混合;溢价倍数延续,但从前沿模型狂热中压缩。$10B-$15B;约为据报道 2026 年 7 月估值的 ~0.6x 至 ~0.8x。收入质量混杂、支持摩擦未解、治理不透明,使折价继续存在。当前证据最能支撑的路径。
悲观ARR 停在当前附近,或客户质量走弱;监管、应用分发或算力约束恶化;剥离后仍在战略上依赖 Kuaishou。$6B-$10B;约为据报道 2026 年 7 月估值的 ~0.3x 至 ~0.6x。倍数压缩和信任恶化可能同时发生。实质下行情景,不是尾部幻想。

区间用于围绕据报道 2026 年 7 月价格约束入场纪律,不是假装当前公开证据足以做精确 DCF 估值。

[CV022, CV023, CV024, CV025, CV026, CV027]
投资逻辑破坏与叫停触发项表
触发项阈值对投资逻辑的传导行动含义
合规 / 应用商店事件任何牵涉 AI 标识或内容治理的正式执法、审核失败、暂停或更新受阻。监管风险从抽象折价变成运营中断。暂停投资判断,直到理解控制体系和分发受损程度。
增长质量恶化相比 Q1 / ARR 轨迹出现可见放缓,且客户质量或企业深度没有相应改善。削弱溢价倍数的核心理由。重切估值,并下调信心。
客户信任侵蚀围绕点数、退款、支持或作品消失的公开投诉持续恶化。指向低质量变现,也说明留存弱于表面用户规模暗示。下调收入质量假设。
剥离后治理不透明2026 年 7 月融资后,独立治理仍没有实质澄清。当前价格下,少数股投资判断和退出规划更难。要求治理折价,或避免入场。
算力 / 出口管制收紧实质性先进芯片新限制,或可见的容量瓶颈。可能同时压缩产品速度和利润率。重做增长、capex / opex 和时点假设。

叫停触发项应能通过公开或尽调证据观察,而不是只靠叙事。

[CV019, CV031, CV037, CV038, CV039]
FV003: 估值 / 回报区间

相对于据报道 2026 年 7 月融资价格的悲观、基准和乐观估值带。

区间用于约束以据报道 2026 年 7 月融资为参照的入场纪律,不是精确的模型估值输出。乐观情景已经假设公司显著执行成功,因此当前入场价的上行空间有限。

[CV022, CV023, CV024, CV025, CV026, CV027]

8.4 退出准备度、尽调要求与最终观点

最终判断因此对价格和证据都很敏感。Kling 已有足够牵引力、产品质量和战略投资人兴趣,从原则上仍具备可投性。缺的是按当前估值把公司当成可持续独立主体承销所需的材料包。投资人仍需要看清 7 月轮后股权结构表和少数股东权利,拿到经审计的独立财务、收入结构与留存细节、企业客户集中度、客服与退款 KPI,以及算力依赖和合规运营的具体图景。单看公开记录,退出故事也还没到可以递表的程度。未来 IPO 或更大的跨轮融资,如果 ARR 快速复合、风险包装改善,有可能在或高于今天估值成交;但今天的公开材料更像高质量动量证据,还不是完整承销证明。因此最终姿态仍是观察。比起又一条普通产品公告,更低的入场估值,或经济性与治理披露大幅改善,会更快改变判断。[CV033, CV034, CV035, CV036, CV037, CV038]

最终尽调清单表
主题缺失证据重要性负责人或尽调路径
股权结构表和治理董事会构成、保留事项、Kuaishou 控制权、投资人保护和剥离分离里程碑。没有治理清晰度,少数股投资无法干净定价。向公司 / 领投方索取股东和治理摘要。
独立财务质量审计收入、毛利率、烧钱速度、现金跑道,以及消费者与企业组合。当前价格需要的不只是标题式 ARR 和季度收入快照。索取审计级财务包和董事会材料。
留存与队列质量NRR / GRR、队列留存、退款 / 拒付率,以及按渠道拆分的使用到付费转化。判断 ARR 是耐久,还是主要靠促销堆出来。索取队列材料和计费质量指标。
企业管线深度具名管线、赢率、扩张队列和大客户集中度。当前公开验证很强,但有些集中在标杆案例上。访谈 GTM 负责人和背书客户。
合规与保证材料包AI 标识架构、备案状态、隐私控制、信任 / 安全材料和事件历史。用于判断监管和采购风险应享有多大折价。索取合规备忘录、信任材料包和发布控制文件。
算力敞口GPU / 云供应商组合、地域部署,以及应对出口政策变化的预案。隐蔽基础设施瓶颈可能同时改变增长和利润率。索取基础设施架构和供应商集中度分析。

这些问题是把“有意思,但不要追价”推进到可投资、价格敏感 yes/no 决策所需的最低集合。

[CV035, CV036, CV040, CV041, CV042]
FV004: 投资 KPI

以 IC 口径,从市场、验证、经济性、治理、风险和估值支撑给 Kling 打分。

[CV005, CV018, CV020, CV028, CV031, CV042]

免责声明

本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;任何投资决策前,都应直接向管理层及原始文件核验。

证据索引

结论
编号陈述可信度来源
CO001 Kling AI is Kuaishou Technology’s publicly marketed AI video and image generation platform. SO001, SO009
CO002 Kuaishou’s disclosed principal place of business in the PRC is Beijing, anchoring Kling’s operating base to the parent’s Beijing footprint. SO009, SO011
CO003 Kuaishou’s public materials state that Kling AI launched in June 2024. SO015, SO016
CO004 By the 2026-08-17 run date, Kling had been in market for roughly 26 months since its June 2024 launch. SO015, SO016
CO005 The Kling 3.0 model family includes Video 3.0, Video 3.0 Omni, Image 3.0, and Image 3.0 Omni. SO015, SO016
CO006 Kuaishou announced the Kling 3.0 model family on February 5, 2026. SO015, SO016
CO007 Kling 3.0 is described as an all-in-one multimodal workflow spanning text, image, audio, and video input and output. SO015, SO016
CO008 Kling Video 3.0 supports video generation up to 15 seconds. SO007, SO016
CO009 Kling 3.0’s native audio generation supports multiple languages, dialects, and accents. SO007, SO016
CO010 Kling officially launched native 4K video generation for the Kling 3.0 series on April 23, 2026. SO003
CO011 Kling Image 3.0 Omni publicly states support for direct 2K and 4K output. SO008
CO012 Kling VIDEO 3.0 Multi-Shot supports up to six shots in one generated sequence. SO007
CO013 Kling’s official August 2026 blog pricing guide lists a Standard plan at $6.99 per month with 660 monthly credits. SO002
CO014 Kling’s official August 2026 blog pricing guide lists a Pro plan at $25.99 per month with 3,000 monthly credits. SO002
CO015 Kling’s official August 2026 blog pricing guide lists a Premier plan at $64.99 per month with 8,000 monthly credits. SO002
CO016 Kling’s official August 2026 blog pricing guide lists an Ultra plan at $127.99 per month with 26,000 monthly credits. SO002
CO017 Kling’s official August 2026 blog pricing guide says native 4K generation should be budgeted at 30 credits per second. SO002
CO018 The fetched U.S. App Store listing shows in-app purchase anchors including Standard monthly at $10, Pro monthly at $37, and Premier monthly at $92. SO006
CO019 The fetched U.S. App Store listing identifies the seller as KLING AI PTE. LTD. SO006
CO020 The fetched U.S. App Store listing shows a 4.7 out of 5 rating based on 28K ratings. SO006
CO021 Kuaishou said Kling exceeded $20 million in monthly revenue in December 2025, implying a $240 million annualized revenue run rate. SO013, SO014
CO022 Kuaishou said Kling had already reached a $100 million ARR milestone in March 2025, around ten months after launch. SO013, SO014
CO023 Kuaishou said Kling served more than 60 million creators worldwide by December 2025. SO013, SO014
CO024 Kuaishou said Kling had generated more than 600 million videos and established partnerships with more than 30,000 enterprise users by late 2025 to early 2026. SO015, SO016
CO025 Kuaishou reported that Kling generated more than RMB650 million of revenue in the first quarter of 2026, representing more than 300% year-over-year growth. SO017, SO018
CO026 Kuaishou reported that Kling’s annualized revenue run rate was approximately USD500 million in March 2026. SO017, SO018
CO027 Kuaishou said Kling launched a Team Plan in the first quarter of 2026 supporting real-time collaborative creation for up to 15 members. SO017, SO018
CO028 Kuaishou said Kling’s Baseball Live effect helped the product reach the number-one App Store position across 42 countries and regions including Brazil and Germany. SO017, SO018
CO029 Kuaishou cited Kling’s use in the Chinese historical drama Swords Into Plowshares and the Hollywood series House of David as evidence of professional production adoption. SO017, SO018
CO030 Kuaishou’s 2025 annual report says growth in the company’s other-services revenue was partly attributable to Kling AI business growth. SO011, SO017
CO031 Multiple July 2026 reports say Kling raised nearly $3 billion, or more than 19 billion yuan, in its spinout financing round. SO019, SO020, SO021, SO022
CO032 Multiple July 2026 reports say Kling’s post-money valuation was about $18 billion. SO019, SO021, SO022, SO026
CO033 Public July 2026 coverage places Kling’s pre-money valuation at about $15 billion. SO020, SO021, SO022
CO034 Public coverage names CPE Yuanfeng, Guofang Venture Capital, BlueFive Capital, Tencent, CITIC Securities, Zhongguancun Science City Fund, and CAS Investment among the lead investors in the July 2026 round. SO019, SO021, SO026
CO035 Public coverage says Alibaba Cloud, Baidu, and Huace Film & TV participated in the July 2026 financing. SO019, SO021, SO026
CO036 CNBC and other July 2026 coverage say Tencent invested about $200 million in the round. SO020, SO021, SO026
CO037 TechNode and MLQ both say the financing was intended to support Kling’s transition to independent commercial operations. SO019, SO026
CO038 Business20 and MLQ both report that Kuaishou retained roughly 68.33% control of Kling after the financing. SO021, SO026
CO039 Cheng Yixiao is disclosed as Kuaishou’s co-founder, chairman of the board, and chief executive officer. SO011, SO023
CO040 Su Hua is disclosed as a Kuaishou co-founder and executive director. SO011, SO023
CO041 Kuaishou’s 2025 annual report says Lu Rong was appointed an independent non-executive director on April 28, 2025. SO011
CO042 The reviewed public source set does not provide a stand-alone Kling AI board page or complete management roster. SO001, SO009, SO010, SO012, SO023
CO043 Kling’s own blog says its native 4K rollout is already being used by film, animation, advertising, and AI-production workflow partners. SO003
CO044 Kuaishou’s 2025 annual report describes Kling as maintaining a globally leading position in multimodal video generation while accelerating commercialization. SO011, SO017
CO045 China’s Measures for Labeling of AI-Generated Synthetic Content apply to AI-generated video and virtual-scene outputs and took effect on September 1, 2025. SO024, SO025
CO046 The labeling measures require both explicit labels visible to users and implicit labels embedded in metadata or watermarks for generated video content. SO024, SO025
CO047 The labeling measures say service providers must provide labeling-related materials during algorithm filing and security-assessment procedures. SO024, SO025
CO048 The labeling measures explicitly make app-distribution platforms verify whether AI-generation apps provide compliant labeling controls during listing review. SO024, SO025
CO049 Kling’s July 2026 blog says The RealReal’s Cannes Lions-winning L’Ultimo Uomo Reale used Kling to maintain believable performances and stable character identity across shots. SO004
CO050 Public July 2026 coverage differs slightly on whether Kling’s financing should be read as just over RMB19 billion raised or a larger capped commitment around RMB20.45 billion. SO021, SO022, SO026
CO051 Public sources reviewed still do not disclose Kling’s stand-alone cap-table detail, liquidation preferences, debt facilities, or a complete public board roster. SO019, SO021, SO026
CM001 Kling AI sells into a narrower AI-video workflow market rather than the full generative-AI economy. SM001, SM002, SM003, SM025
CM002 Kling's visible commercial surfaces include creator subscriptions, mobile app usage, enterprise APIs, and production-oriented video workflows. SM001, SM005, SM006, SM025
CM003 The company's addressable marketing and e-commerce spend includes branded short-video and product-demo creation budgets. SM003, SM025
CM004 Kling also targets film and television workflows through storyboard, VFX, and synthetic-scene use cases. SM003, SM027
CM005 Kling's API surface makes developer and product budgets part of its serviceable market. SM001, SM023
CM006 Text-only copilots, generic enterprise search, and GPU hardware revenue should be excluded from Kling-specific market sizing. SM001, SM007, SM008
CM007 Broad generative-AI TAM charts overstate Kling's nearer-term serviceable market if used without workflow-level boundary logic. SM007, SM008, SM009
CM008 MarketsandMarkets estimates the global generative AI market at USD 185.45 billion in 2026. SM007
CM009 MarketsandMarkets projects the global generative AI market to reach USD 1,658.97 billion by 2033 at a 36.8% CAGR. SM007
CM010 Statista describes the generative AI market as growing at more than 24.4% CAGR in the coming years. SM008
CM011 Adwave summarizes Grand View Research as valuing the AI video generator market at about USD 788.5 million in 2025. SM009
CM012 Adwave summarizes Grand View Research as projecting the AI video generator market to about USD 3.44 billion by 2033 at a 20.3% CAGR. SM009
CM013 Kuaishou disclosed that Kling AI generated over RMB650 million of revenue in Q1 2026. SM003, SM026
CM014 Kuaishou disclosed that Kling AI's March 2026 annualized revenue run rate was approximately USD 500 million. SM003, SM026
CM015 Kuaishou disclosed that Kling AI served over 60 million creators worldwide as of December 2025. SM002, SM027
CM016 Kuaishou disclosed that Kling AI had generated more than 600 million videos as of December 2025. SM002, SM027
CM017 Kuaishou disclosed that Kling AI had established partnerships with over 30,000 enterprise users by December 2025. SM002, SM027
CM018 Contradictory market estimates should be preserved because broad GenAI and dedicated AI-video studies measure different layers of demand. SM007, SM008, SM009
CM019 Individual creators are a primary buyer class for Kling AI. SM001, SM005, SM006
CM020 Marketing and e-commerce teams are a primary buyer class for Kling AI. SM003, SM025
CM021 Film, television, and premium creative teams are a primary buyer class for Kling AI. SM003, SM027
CM022 Developer and product teams are a primary buyer class for Kling AI where video generation is embedded through APIs. SM001, SM023
CM023 In creator tiers, the Kling buyer, user, and payer often collapse into one individual account. SM005, SM006
CM024 In marketing deployments, the day-to-day users are often creators or marketers while the payer is a brand, agency, or commerce budget owner. SM003, SM025
CM025 In film and studio use cases, the buyer or payer is more likely to sit in a production budget than in consumer software spend. SM003
CM026 In API deployments, integration ownership is likely to sit with product, platform, or IT budget owners rather than creator budgets. SM001, SM023
CM027 Marketing and e-commerce adoption is triggered primarily by content velocity, lower production cost, and asset consistency. SM009, SM025
CM028 Film and premium creative adoption is triggered primarily by controllability, character consistency, and cinematic motion. SM003, SM010, SM023
CM029 Mobile app distribution gives Kling a shorter trial path than pure API vendors because users can discover, create, and subscribe in one interface. SM005, SM006
CM030 The presence of community browsing, cloning, and remix features supports a creator-led adoption funnel rather than a purely enterprise-sales motion. SM005, SM006
CM031 Coherent describes the 2026 AI-video market as shifting from novelty toward controllable creative infrastructure. SM010
CM032 Repeatable control, editing flexibility, API integration, and production-ready output are becoming more important than one-off visual spectacle. SM010
CM033 China's AI-generated-content labeling rules raise compliance and provenance requirements for public-facing AI video products. SM011, SM012
CM034 China Law Translate highlights that app-distribution platforms must verify labeling measures during app listing checks. SM011, SM012
CM035 Global expansion of AI video products also faces broader cross-border regulatory and disclosure friction. SM011
CM036 Sora's 2026 shutdown is a caution that AI-video demand does not guarantee sustainable unit economics. SM013
CM037 DigitalApplied frames the AI-video sector as growing quickly after Sora's shutdown rather than collapsing with it. SM013
CM038 Price compression and category competition make retention, workflow fit, and margin structure as important as raw model quality. SM009, SM010, SM013, SM015, SM020, SM022
CM039 Runway, Veo, Pika, Luma, Wan, and other rivals give buyers many ways to multi-home across quality, price, and ecosystem preferences. SM014, SM015, SM016, SM017, SM018, SM019, SM020, SM021, SM022, SM024
CM040 The most defensible concise statement of Kling's market is programmable short-form visual generation across creator, marketing, e-commerce, studio, and API workflows. SM001, SM003, SM005, SM025
CM041 Alibaba-backed Wan adds competitive pressure by expanding the supply of advanced video-model alternatives in 2026. SM024
CM042 Public sources do not cleanly isolate Kling's SAM by geography, vertical, or creator-versus-enterprise revenue mix. SM003, SM004
CP001 Kling competes across creator apps, production workflows, and enterprise/API video generation rather than in a single narrow lane. SP001, SP002, SP003
CP002 The correct competitor set changes by buyer job rather than by one universal product category. SP001, SP004, SP007, SP009, SP012, SP013
CP003 Runway is a direct rival for professional teams that want generation and editing in one workspace. SP004, SP006
CP004 Google Veo is a direct rival for premium cinematic and enterprise use cases. SP007, SP008
CP005 Pika is a direct rival in creator and social-first video creation. SP009, SP010
CP006 Kling's strongest public fit is cinematic short-form output with believable human motion and character continuity at accessible price points. SP013, SP015, SP022, SP025
CP007 China-linked buyers and cost-sensitive teams also face alternative supply from Wan, Hailuo, Seedance, and routing layers. SP013, SP016, SP020
CP008 Status quo and internal production remain substitutes where vendor risk or creative control concerns outweigh AI speed gains. SP004, SP012, SP021
CP009 Kling's public proof of scale differentiates it from purely experimental rivals because Kuaishou discloses large creator and enterprise adoption. SP002, SP003
CP010 Runway positions itself as an all-in-one cloud creative platform with video, image, and audio generation plus editing. SP004
CP011 Runway Gen-4 emphasizes consistent characters, locations, and objects across scenes. SP006
CP012 Veo 3 emphasizes native audio, realism, prompt adherence, and creative control. SP008
CP013 Veo benefits from the broader Google ecosystem rather than only from standalone model quality. SP007, SP008
CP014 Pika is positioned around accessible video creation and creator-friendly experimentation. SP009, SP010
CP015 Independent comparison coverage repeatedly highlights Kling for strong human motion and character continuity. SP013, SP015
CP016 Independent comparison coverage highlights Runway for camera control and professional workflow depth. SP014, SP015, SP016
CP017 Independent comparison coverage highlights Veo for native audio and premium realism. SP014, SP016
CP018 Luma positions itself around creative agents, collaboration, and production delivery rather than only raw clip generation. SP011, SP012
CP019 Luma's official surfaces emphasize team workspaces, shared context, and export for production. SP011, SP012
CP020 No single retrieved source proves a universal quality winner, which reinforces that buyer fit is workload-specific. SP014, SP015, SP016
CP021 Kling's August 2026 blog guide lists consumer plans from $6.99 standard to $127.99 ultra. SP022
CP022 The fetched App Store listing shows higher in-app purchase anchors such as $37 monthly for Pro, indicating channel variance in consumer pricing. SP025
CP023 CostBench summarizes Kling API pricing at roughly $0.084 to $0.420 per second depending on mode and quality. SP018
CP024 WaveSpeed argues that actual Kling API economics depend on seconds, resolution, audio, retries, and accepted-output rate rather than list price alone. SP017
CP025 Independent comparison coverage places Runway above Kling on premium workflow depth but at higher subscription price points. SP015, SP016
CP026 Independent comparison coverage places Pika below Kling on photorealistic production but at low-friction creator price points. SP014, SP015
CP027 Veo often competes through ecosystem access and premium capability rather than a simple studio-style list-price comparison. SP007, SP008, SP016
CP028 Aggregation layers and programmatic access make AI-video pricing highly task-specific and reduce the relevance of one canonical vendor price. SP017, SP018
CP029 Many teams can rationally use multiple tools in one pipeline, such as one for hero shots and another for high-volume variants. SP015, SP017
CP030 Low switching cost is a central competitive fact in AI video because routing layers and modular workflows reduce single-vendor dependence. SP012, SP017, SP018
CP031 Kuaishou backing gives Kling capital, distribution adjacency, and proof of rapid commercialization. SP002, SP003
CP032 Kling's public adoption metrics show it is far beyond demo-stage relevance. SP002, SP003
CP033 Premium rivals such as Runway and Veo are strong enough on workflow and ecosystem power to threaten Kling in enterprise and film use cases. SP004, SP006, SP007, SP008, SP016
CP034 Creator-first rivals such as Pika and workflow systems such as Luma can compete for lighter-weight or team-oriented workloads even without matching Kling on every quality metric. SP009, SP010, SP011, SP012, SP014
CP035 Wan and other Chinese alternatives increase competitive pressure by expanding the available supply of capable video models. SP013, SP020
CP036 Sora's discontinuation shows that category visibility does not itself create durable economics or defensibility. SP019, SP021
CP037 Kling's moat is more proven on quality-per-dollar and commercialization than on customer exclusivity or switching cost. SP002, SP003, SP017, SP018
CP038 Public sources do not prove durable enterprise switching costs for Kling specifically. SP002, SP003, SP017
CP039 Public sources do not prove that Kling customers use the product exclusively rather than alongside other video vendors. SP017, SP018
CP040 The most defensible concise conclusion is that Kling belongs in the leading AI-video cohort but still operates in a market with weak lock-in and strong multi-homing behavior. SP013, SP015, SP016, SP017, SP018, SP021
CI001 Kling monetizes through multiple visible surfaces including memberships, in-app purchases, API usage, enterprise packages, and team-oriented workflow features. SI001, SI002, SI003, SI012
CI002 Consumer memberships are a formal recurring revenue surface for Kling. SI001, SI005
CI003 In-app purchases are a formal monetization surface for Kling via app-store channels. SI002
CI004 Usage-based API pricing and prepaid packages indicate a developer and enterprise monetization surface beyond consumer subscriptions. SI003, SI004, SI021
CI005 Kling publicly supports team collaboration, implying a workflow upsell beyond one-person creation. SI012
CI006 Public sources do not disclose what share of Kling revenue comes from each monetization surface. SI009, SI010, SI011
CI007 Kling’s August 2026 blog guide lists a Standard membership at $6.99 per month with 660 credits. SI001
CI008 Kling’s August 2026 blog guide lists a Pro membership at $25.99 per month with 3,000 credits. SI001
CI009 Kling’s August 2026 blog guide lists a Premier membership at $64.99 per month with 8,000 credits. SI001
CI010 Kling’s August 2026 blog guide lists an Ultra membership at $127.99 per month with 26,000 credits. SI001
CI011 The fetched App Store listing shows channel pricing anchors of $10 monthly for Standard, $37 for Pro, and $92 for Premier. SI002
CI012 CostBench summarizes Kling API pricing at roughly $0.084 to $0.420 per second depending on mode and quality. SI003
CI013 Kling paid-service terms define credits as non-cash virtual tools rather than stored monetary value. SI005, SI023
CI014 Kling paid-service terms state that purchased credits have a validity period of two years. SI005, SI023
CI015 Kling’s official pricing guide uses 30 credits per second as the 4K planning reference. SI001
CI016 WaveSpeed argues that true Kling API spend should be modeled as cost per usable video rather than cost per request. SI004, SI022
CI017 WaveSpeed highlights retries, rejected outputs, and review time as important hidden costs in Kling API deployment. SI004, SI022
CI018 Kling paid-service terms state that refunds or reverse exchanges for credits are not supported. SI005
CI019 Public sources do not disclose Kling gross margin by mode, workflow, or customer segment. SI009, SI010, SI011
CI020 Public sources do not disclose Kling monthly burn or standalone runway after the spin-out financing. SI010, SI013, SI014
CI021 Kuaishou disclosed December 2025 monthly revenue above USD20 million for Kling, implying USD240 million ARR. SI009
CI022 Kuaishou disclosed Q1 2026 Kling revenue above RMB650 million. SI010, SI011
CI023 Kuaishou disclosed March 2026 Kling ARR of approximately USD500 million. SI010, SI011
CI024 Kuaishou described Kling Q1 2026 revenue growth as more than 300% year over year. SI010
CI025 Kuaishou reported that growth in its other services segment was primarily due to Kling AI business growth in Q1 2026. SI010
CI026 Kling had already reached material commercial scale before the July 2026 spin-out financing. SI009, SI010, SI011
CI027 Independent coverage places Kling’s July 2026 financing at roughly USD2.8 billion to nearly USD3.0 billion. SI014, SI016, SI017, SI018, SI019
CI028 Independent coverage places Kling’s July 2026 post-money valuation at about USD18 billion. SI014, SI018, SI019
CI029 Independent coverage says Kuaishou retained roughly 68.33% control after the July 2026 transaction. SI014, SI017, SI019
CI030 Kuaishou’s 2025 annual report states total available funds of RMB117.7 billion as of March 31, 2026. SI013
CI031 Kuaishou parent liquidity directionally lowers near-term capital risk for Kling even though it is not the same as standalone cash. SI013, SI014, SI016
CI032 Public sources do not disclose Kling standalone cash on hand or exact use-of-funds allocation after the July 2026 financing. SI014, SI016, SI018
CI033 Trustpilot review aggregation presents Kling with a 1.3/5 rating and frequent mentions of subscriptions, refunds, payment, and customer service. SI020
CI034 Adverse Trustpilot reviews describe wasted credits, refund denials, subscription-management confusion, and poor support responsiveness. SI020
CI035 Kling’s terms acknowledge that outputs may not always be accurate, which can increase costly retries for quality-sensitive work. SI006
CI036 The paid-service terms allow the company to update pricing schemes and related benefits over time. SI005
CI037 Public sources do not isolate Kling revenue by geography or by consumer versus enterprise segment. SI009, SI010, SI011
CI038 Kling’s financial profile could differ materially depending on whether enterprise/API or consumer/app-store channels dominate revenue. SI001, SI002, SI003, SI012
CI039 Sora’s discontinuation is a category caution that growth in AI video does not itself prove healthy unit economics. SI025
CI040 The most defensible financial verdict is that Kling has proven top-line demand and capital access, but still lacks public proof on margin quality, burn efficiency, and retention-driven revenue durability. SI009, SI010, SI011, SI014, SI020, SI025
CI041 Kling maintains a live pricing page, reinforcing that current list pricing can change outside static blog explanations. SI026, SI005
CE001 Kling’s public product stack includes Video 3.0, Video 3.0 Omni, Image 3.0, and Image 3.0 Omni. SE005, SE009
CE002 Kling also has visible web, mobile, community, and API surfaces rather than only one studio interface. SE001, SE003, SE004, SE017, SE018
CE003 Kling’s public materials frame the product as an all-in-one creative engine rather than only a text-to-video tool. SE001, SE005, SE009
CE004 Creator ideation, marketing production, e-commerce visuals, and film/storyboard workflows are all explicitly supported use cases in public materials. SE013, SE014, SE017, SE025
CE005 The API page and release-note surfaces indicate that Kling treats programmability and product iteration as part of the product story. SE001, SE004, SE005, SE006
CE006 The app and community surfaces indicate that Kling also treats discovery, remixing, and creator inspiration as part of the product stack. SE003, SE017, SE018
CE007 Kling public materials support text-to-video, image-to-video, and reference-to-video workflows. SE005, SE009
CE008 Kling public materials support editing-style operations such as adding, removing, modifying, or transforming video content. SE005
CE009 The app-store listing shows Kling supports video extension up to three minutes on consumer surfaces. SE017
CE010 Kling 3.0 is publicly described as a unified multimodal model spanning text, image, audio, and video input/output. SE001, SE005, SE009
CE011 Kling public release notes describe the architecture as integrating multiple tasks inside one native multimodal training model. SE005
CE012 Multi-Shot is designed to turn scene coverage and shot structure into one generated cinematic sequence. SE005, SE010
CE013 Kling VIDEO 3.0 supports subject consistency through multi-image references and even video references as elements. SE005
CE014 Kling public materials say Native Audio now supports more languages, dialects, accents, and multi-character dialogue scenes. SE005, SE009
CE015 Kling public materials say Video 3.0 supports generation up to 15 seconds with flexible duration from 3 to 15 seconds. SE005, SE011, SE009
CE016 The 4K product guidance indicates that Kling uses native 4K output as a premium production mode rather than mere upscaling. SE012, SE013
CE017 IMAGE 3.0 Omni is publicly described as supporting direct 2K and 4K output for professional visual work. SE006
CE018 IMAGE 3.0 technical notes describe a Visual Chain-of-Thought style inference approach and other cinematic-quality upgrades. SE006, SE016
CE019 Public materials do not expose Kling’s full internal serving topology, but they do expose meaningful operating-layer detail through release notes. SE005, SE006, SE016
CE020 Kling has visible release notes, resource pages, and an API page, which are signs of a product that is iterating in public. SE001, SE004, SE005, SE006
CE021 Kuaishou disclosed a Team Plan supporting real-time collaborative creation for up to 15 members. SE015
CE022 The e-commerce 4K workflow guide describes real production choices such as aspect ratio, clip duration, camera moves, and quality checks. SE013
CE023 The L’Ultimo Uomo Reale write-up frames Kling as a tool capable of maintaining stable character identity and cinematic consistency across production. SE014, SE024, SE025
CE024 Independent film-oriented coverage shows Kling is being discussed in real production contexts rather than only consumer creation contexts. SE024, SE025
CE025 Kling’s release cadence from O1/2.6 to 3.0 and 4K rollout indicates rapid product iteration. SE009, SE012, SE015
CE026 Public sources still leave gaps on uptime history, SLAs, enterprise admin controls, and support boundaries. SE001, SE007, SE008
CE027 The most visible maturity proof today is breadth of workflow surfaces and release cadence, not formal enterprise-control disclosure. SE001, SE004, SE005, SE006, SE015
CE028 Kling’s public terms warn that outputs may not always be accurate. SE007
CE029 Kling’s public terms warn that services are provided on an as-is and as-available basis and may become unavailable. SE007
CE030 Kling’s public terms require users to hold rights to input content and to review outputs for appropriateness. SE007
CE031 China’s labeling rules require conspicuous labeling and implicit metadata-based labeling for AI-generated video content. SE019, SE020
CE032 China Law Translate highlights that app distribution platforms must verify labeling measures during listing checks. SE020
CE033 Kling’s privacy policy says the service stores user data on servers in Singapore and processes uploaded and generated content broadly. SE008
CE034 Kling’s privacy policy says the service may process images, audio, and generated content for moderation and operations. SE008
CE035 The paid-service terms say pricing, services, and benefits can be updated on front-end pages over time. SE021
CE036 The app-store and public legal surfaces prove consumer distribution and governance exist, but do not prove full enterprise procurement readiness. SE017, SE018, SE007, SE008
CE037 Independent comparison sources repeatedly highlight Kling for motion realism, facial animation, and character continuity. SE022, SE023
CE038 Public product evidence is stronger on creative differentiation than on formal control and governance transparency. SE005, SE006, SE007, SE008, SE021
CE039 Public sources do not prove enterprise-grade security certifications or formal SLAs for Kling. SE007, SE008
CE040 The most defensible product-tech verdict is that Kling has a sophisticated, commercially oriented multimodal product stack with meaningful workflow differentiation, but still requires deeper diligence on reliability, governance, and enterprise controls. SE001, SE005, SE006, SE007, SE008, SE015, SE022, SE023
CU001 Kling’s visible customer base spans self-serve creators, film and television teams, brand/agency marketers, enterprise workgroups, and a smaller developer/product-team surface. SU002, SU003, SU005, SU006, SU010, SU011
CU002 Buyer user and payer are often the same person in consumer creator usage but split across producer team lead and operator roles in film agency and enterprise contexts. SU002, SU014, SU015, SU022
CU003 Kling has a genuinely global creator surface rather than a China-only user base. SU002, SU003, SU005, SU006
CU004 Film and television production is one of Kling’s clearest publicly proven customer segments. SU002, SU008, SU014, SU015, SU018
CU005 Marketing advertising and e-commerce teams are also a visible commercial segment for Kling. SU004, SU007, SU008, SU009, SU016
CU006 The Team Plan and the 30,000-enterprise-user disclosure indicate a workgroup or enterprise layer beyond individual creator subscriptions. SU001, SU002, SU003
CU007 App stores the web studio and project publicity are the main publicly visible acquisition surfaces for Kling. SU005, SU006, SU018, SU025
CU008 Kling relies on mobile storefronts as important consumer discovery and billing rails. SU005, SU006
CU009 Kuaishou’s Q1 disclosure that Kling reached the number one App Store position across 42 countries and regions shows acquisition bursts can be globally distributed. SU002
CU010 Public channel proof is stronger than public customer-quality proof because distribution surfaces are visible while revenue and cohort disclosures are not. SU005, SU006, SU022, SU023
CU011 Kling served more than 60 million creators worldwide by February 2026. SU001, SU003
CU012 Kling had generated more than 600 million videos by the December 2025 / February 2026 disclosure set. SU001, SU003
CU013 Kling had partnerships with or enterprise usage from more than 30,000 enterprise users or clients by the same disclosure set. SU001, SU003
CU014 Kling’s consumer acquisition surged strongly enough in Q1 2026 to top the App Store across 42 countries and regions. SU002
CU015 The U.S. App Store provided a substantial live customer signal on the run date with a 4.7 rating from 28,000 ratings. SU005
CU016 Google Play provided an even larger Android footprint on the run date with 10 million-plus downloads and more than 528,000 reviews. SU006
CU017 House of David is a strong named production proof because Variety documented 72 AI-assisted shots in season one and later Cannes coverage said Kling generated the majority of the show’s AI production shots. SU014, SU018, SU019
CU018 House of David also provides early repeat-usage evidence because public sources describe Kling use across both season one and season two. SU019
CU019 Swords Into Plowshares is a meaningful production proof because Timeaxis said Kling was integrated across the pipeline from previs through final compositing plates. SU002, SU015
CU020 The RealReal’s L'Ultimo Uomo Reale is a credible brand-workflow proof because both Kling and independent sources tie the campaign to Kling-enabled character consistency and Cannes recognition. SU007, SU016, SU017, SU020
CU021 Raphael and MINIBOTS show that Kling is pursuing international feature-film and animation relationships beyond one-off China use cases. SU018, SU019
CU022 Most named customer proof demonstrates concrete production use or partnership activity but not contract value renewal or long-term account economics. SU014, SU015, SU018, SU019
CU023 Kling’s public customer record is stronger on production references and creator-scale metrics than on classic enterprise case-study detail. SU001, SU014, SU015, SU018
CU024 Curated creator quotes and project publicity are helpful acquisition and positioning signals but are weaker than audited customer outcome disclosures. SU005, SU006, SU025
CU025 No retained public source discloses Kling NRR GRR logo churn or standard cohort-retention metrics. SU001, SU002, SU005, SU006
CU026 Kling monetizes creators through memberships and credits that are explicitly governed by recurring paid-service terms. SU005, SU022
CU027 Activated paid services are non-refundable and credits are governed as virtual tools rather than stored monetary value. SU022
CU028 Kling’s terms permit continuing automatic renewal until explicitly canceled and reserve the ability to adjust prices and benefits. SU022
CU029 The service agreement does not promise uninterrupted availability and instead offers the service on an as-is and as-available basis. SU023
CU030 Privacy and service terms show that Kling processes uploaded and generated content and reserves meaningful rights around content handling which can slow enterprise procurement. SU023, SU024
CU031 Trustpilot preserves a material adverse signal around billing support and wasted credits with a visible 1.3/5 overall rating. SU021
CU032 Google Play reviews show mixed but concrete user sentiment including praise for image-to-video quality alongside complaints about credits and disappearing outputs. SU006
CU033 Billing refund and support quality are therefore visible customer risks even if underlying generation quality can be strong. SU006, SU021, SU022
CU034 For serious professional users legal and support friction can matter almost as much as creative quality because credit burn and content-rights uncertainty directly affect willingness to renew. SU021, SU022, SU023
CU035 App-store adoption and ratings prove live usage but they do not prove paid retention or enterprise-quality durability. SU005, SU006
CU036 Kling has a plausible land-and-expand path from creator experimentation into paid subscriptions then into team plans and higher-value production work. SU002, SU005, SU006, SU022
CU037 Named film and brand projects suggest that one successful project can expand into broader studio or agency workflow adoption. SU007, SU008, SU014, SU019
CU038 Kling’s official materials explicitly target e-commerce and marketing production which broadens expansion potential beyond filmmakers. SU004, SU009
CU039 The sharper public concentration risk is channel dependence on app stores Kuaishou-backed narrative distribution and marquee references rather than any single disclosed whale customer. SU002, SU005, SU006, SU018
CU040 Public sources do not reveal customer or revenue concentration by geography platform consumer-vs-enterprise mix or top account. SU001, SU002, SU005, SU006
CU041 The most supportable customer verdict is that Kling has real global adoption breadth and premium-workflow credibility but still lacks the public durability data needed for clean customer-quality underwriting. SU001, SU002, SU014, SU021, SU022
CR001 China’s 2025 labeling measures require explicit and implicit identification for AI-generated text, image, audio, video, and virtual scenes. SR001, SR002, SR005
CR002 Those rules explicitly extend to downloaded, copied, or exported generated files, not just on-screen display. SR001, SR002
CR003 China app-distribution platforms must verify labeling-related materials for apps offering generative-AI services. SR001, SR002, SR005
CR004 China’s broader generative-AI regime includes filing or registration obligations for public-facing services and applications. SR003, SR004, SR006
CR005 As of February 2026, hundreds of generative-AI services and applications had completed filing or registration in China, showing the regime is active rather than theoretical. SR003, SR006
CR006 Violations can route through multiple Chinese legal frameworks and agencies rather than one narrow AI-specific penalty channel. SR004, SR005, SR006
CR007 Kling’s service agreement states the service is offered on an as-is and as-available basis, which is a direct reliability and contracting risk. SR008
CR008 Kling’s paid-service terms make activated paid services non-refundable, which can amplify customer-support and dispute risk. SR010
CR009 Kling’s privacy policy confirms broad storage and processing of uploaded and generated content, which raises enterprise procurement questions around data handling. SR009
CR010 Cross-border operators increasingly face overlapping AI, privacy, and transparency regimes rather than one harmonized rulebook. SR004, SR007
CR011 The legal and regulatory stack therefore combines product labeling, app review, filing, privacy, and contract obligations in a way that can directly affect product operations. SR001, SR003, SR008, SR009
CR012 Operational risk is visible because public complaint surfaces repeatedly mention burned credits, disappearing generations, and poor support responsiveness. SR011, SR012
CR013 Trustpilot provides a materially adverse public signal with a 1.3/5 rating and repeated billing/support complaints. SR011
CR014 Google Play reviews show that at least some users perceive strong image-to-video quality but still complain about credits and vanished outputs. SR012
CR015 App-store adoption proof does not offset the risk that refund, chargeback, and support costs could rise with consumer scale. SR012, SR013, SR010
CR016 Apple and Google are not just distribution channels; they are policy and billing gatekeepers for a large share of Kling’s consumer footprint. SR012, SR013
CR017 Public sources show real film and campaign workflows using Kling, which raises the business cost of outages, quality drift, or rights-management failures. SR018, SR019, SR020, SR021
CR018 Kling is trying to serve creators, marketers, and film teams across text, image, audio, and video workflows, which increases moderation and control complexity. SR014, SR016, SR029
CR019 Public evidence remains thin on enterprise-grade SLAs, certifications, or incident history relative to the platform’s commercial ambition. SR008, SR009, SR012
CR020 The trust gap is therefore not product absence but assurance opacity. SR008, SR009, SR019
CR021 Scaling from a viral creator product into a dependable professional platform requires heavier support, invoicing, compliance, and success operations than public materials currently prove. SR010, SR011, SR012, SR020
CR022 Kuaishou remains the controlling parent after the July 2026 round, with outside reporting placing its post-raise stake around 68%. SR022, SR023
CR023 That ownership concentration means the spinout is still economically and strategically tied to Kuaishou even as it moves toward independence. SR022, SR023, SR024
CR024 Kling’s premium-workflow credibility depends heavily on a relatively small set of public marquee references such as House of David, Swords Into Plowshares, and The RealReal. SR018, SR019, SR020, SR021
CR025 App stores remain critical discovery and monetization rails for Kling’s global consumer reach. SR012, SR013, SR014
CR026 Advanced AI-chip access to China remains policy-mediated rather than guaranteed, even after the January 2026 U.S. shift to case-by-case licensing for certain chips. SR025, SR026, SR027
CR027 If Kling’s cost structure or velocity depends on constrained advanced compute, export-control changes can transmit quickly into product and margin risk. SR025, SR026
CR028 Strategic investor overlap can create conflicting incentives because some backers are also major technology ecosystems or competing AI platforms. SR022, SR024
CR029 The public record is still thin on standalone board composition, minority-protection terms, or succession depth for the spinout. SR017, SR022, SR023
CR030 Kling’s execution risk therefore sits as much in operating separation and governance maturity as in raw model quality. SR017, SR022, SR029
CR031 Continual model iteration is itself a dependency, because user expectations and professional proofs rest on visible release cadence and product upgrades. SR014, SR028, SR029, SR030
CR032 The July 2026 financing implies an approximately $18B post-money valuation and therefore a lower tolerance for unresolved governance, retention, and compliance risk. SR023, SR024
CR033 Primary sources show meaningful commercialization — over RMB650M Q1 2026 revenue and about $500M March ARR — but not standalone margin or burn durability. SR014, SR015
CR034 Public primary sources do not disclose standalone Kling gross margin, burn, runway, or consumer-vs-enterprise revenue mix. SR014, SR015, SR017
CR035 Secondary reporting on overseas mix, independence plans, or investor composition is useful but not a substitute for audited segment financials. SR022, SR023, SR024
CR036 Non-refundable credits and shifting plan economics can create a margin-vs-customer-trust tension as usage scales. SR010, SR011, SR012
CR037 Building the compliance, support, and enterprise-assurance stack needed for a standalone company will likely raise operating expense before it reduces risk. SR006, SR007, SR022
CR038 If marquee proof points fail to translate into broader repeatable enterprise accounts, valuation support can weaken even with continued headline adoption. SR018, SR019, SR022
CR039 A formal China compliance failure, app-store disruption, or material deterioration in public support complaints would be an immediate underwriting alarm. SR001, SR011, SR012, SR013
CR040 A renewed tightening in advanced-chip access or visible product-capacity constraints would warrant re-cutting growth and margin assumptions. SR025, SR026, SR027
CR041 Absent clearer standalone governance and customer-quality disclosure, the prudent stance is to use strict entry discipline and ongoing monitoring rather than assume risk is already normalized. SR017, SR022, SR023, SR024
CR042 The most supportable current verdict is that Kling’s risk stack is investable only with heavy diligence on compliance, governance, durability, and compute dependence. SR014, SR017, SR023, SR026
CV001 Independent July 2026 reporting converges around an approximately $2.8B financing at an $18B post-money valuation for Kling AI. SV001, SV002, SV003
CV002 The July 2026 round is framed as a step toward more independent commercial operations rather than proof of full operational independence already achieved. SV002, SV003
CV003 Kuaishou’s Q1 2026 disclosure says Kling’s March 2026 ARR was approximately $500M. SV004
CV004 Kuaishou’s Q1 2026 disclosure says Kling generated over RMB650M of revenue in Q1 2026, up more than 300% year over year. SV004
CV005 Primary and high-quality independent sources together show that Kling has real commercial and professional-workflow adoption rather than only viral consumer interest. SV004, SV005, SV008, SV009
CV006 The current public evidence therefore supports paying attention to Kling as a major company, not dismissing it as speculative noise. SV004, SV005, SV008
CV007 The public record still relies heavily on parent-company disclosures instead of audited standalone financial reporting. SV004, SV007
CV008 That disclosure structure weakens confidence in precise valuation work even when top-line growth looks strong. SV004, SV007
CV009 The most supportable recommendation at the current reported mark is Track rather than Buy. SV001, SV004, SV007
CV010 A better disclosure package or lower price would change the recommendation faster than another general product announcement would. SV007, SV028
CV011 Adobe trades at roughly 4.17x market cap to TTM revenue based on August 2026 CompaniesMarketCap data. SV019, SV020
CV012 Duolingo trades at roughly 5.70x market cap to TTM revenue based on August 2026 CompaniesMarketCap data. SV021, SV022
CV013 GitLab trades at roughly 7.19x market cap to TTM revenue based on August 2026 CompaniesMarketCap data. SV023, SV024
CV014 C3.ai trades at roughly 6.16x market cap to TTM revenue based on August 2026 CompaniesMarketCap data. SV025, SV026
CV015 Those public references cluster around materially lower revenue multiples than Kling’s reported July 2026 valuation implies. SV019, SV020, SV021, SV022, SV023, SV024, SV025, SV026
CV016 Kling likely deserves a premium to mature public software comps because growth is faster and strategic scarcity is higher. SV004, SV005, SV008, SV027, SV028
CV017 But the size of the currently implied premium requires unusually strong proof on governance, economics, and risk control that is not yet public. SV007, SV014, SV018, SV030
CV018 Public customer-friction evidence creates a quality discount even if total adoption is impressive. SV012, SV013, SV014, SV015, SV016
CV019 Non-refundable credits and as-is / as-available service terms can make revenue quality look weaker than gross booking growth alone suggests. SV015, SV016
CV020 Thin public assurance and governance packaging reduce the willingness to pay peak software multiples for Kling today. SV007, SV015, SV017
CV021 Current premium-workflow proof is meaningful but still concentrated in a relatively small set of marquee references. SV008, SV009, SV010, SV011
CV022 Using the disclosed March 2026 ARR anchor, the reported $18B valuation implies roughly a 36.0x ARR multiple. SV001, SV004
CV023 That implied multiple is difficult to justify without very rapid forward ARR growth and a cleaner risk profile. SV004, SV014, SV018
CV024 A bull case requires ARR to surpass about $1B within roughly 12-18 months while governance, compliance, and customer-quality indicators improve. SV004, SV005, SV027, SV028
CV025 A bull valuation range of roughly $18B-$24B is supportable only if several positive assumptions land together. SV004, SV027, SV028
CV026 A base case assumes continued growth but persistent mixed revenue quality, incomplete risk normalization, and partial multiple compression. SV004, SV014, SV015, SV018
CV027 A base valuation range of roughly $10B-$15B is the most supportable path from current public evidence. SV004, SV014, SV019, SV020, SV021, SV022
CV028 A bear case becomes plausible if support friction, regulation, distribution, or compute access worsen before governance and economics are clarified. SV014, SV018, SV029, SV030
CV029 A bear valuation range of roughly $6B-$10B is not a tail fantasy given the current risk stack and the valuation starting point. SV014, SV018, SV030
CV030 At the reported July 2026 entry price, even the bull case offers far less upside asymmetry than the remaining diligence gaps would normally warrant. SV001, SV004, SV007
CV031 The current reported mark therefore has limited margin of safety. SV001, SV004, SV014, SV018
CV032 Investors should prefer lower entry price or better disclosure over paying up simply because the company is clearly growing. SV001, SV004, SV007
CV033 A credible pro-thesis exists: large creator scale, fast release cadence, premium workflow adoption, and major strategic capital all support a significant company outcome. SV001, SV004, SV005, SV008, SV027, SV028
CV034 A credible anti-thesis exists: consumer-friction signals, governance dependence, regulation, and incomplete unit economics can prevent the current price from earning an attractive return. SV014, SV015, SV016, SV018, SV030
CV035 Exit readiness is incomplete on the public record alone because there is no standalone audited package in the retained evidence set. SV007
CV036 The next diligence step should focus on cap table, governance rights, and separation milestones after the July 2026 round. SV002, SV003, SV007
CV037 The next diligence step should also focus on audited financial quality: gross margin, burn, runway, revenue mix, and cohort retention. SV004, SV007, SV016
CV038 A formal China compliance or app-review failure would be an immediate thesis-break event. SV012, SV013, SV018, SV029
CV039 A material tightening in advanced-chip access or visible compute bottlenecks would force a lower valuation range. SV029, SV030
CV040 If growth slows materially without better customer-quality evidence, the main rationale for the premium multiple weakens quickly. SV004, SV014
CV041 If the spinout remains opaque on minority protections and control rights, exit planning and public-market readiness both become harder to underwrite. SV002, SV003, SV007
CV042 The final price-sensitive verdict is Track: Kling is a real and important business, but the current reported valuation is not yet sufficiently supported by public evidence. SV001, SV004, SV007, SV014, SV018
来源
编号出版方标题引文
SO001 Kling AI Kling AI: Next-Gen AI Video & Image Generator All-New KlingAI 3.0 Series — All in One, One for All.
SO002 Kling AI Blog Kling 3.0 Credit Cost: 4K, Omni Audio, and Multi Shot Pricing The 4K release introduced cinema-grade native 4K video output for the Kling VIDEO 3.0 series at 30 credits per second.
SO003 Kling AI Blog Kling AI Introduces Native 4K Video Model On April 23, Kling AI officially launched the world’s first native 4K video generation feature for the Kling 3.0 video model series.
SO004 Kling AI Blog Beyond Realism: How Kling Powered L'Ultimo Uomo Reale, a Cannes Lions-winning Film Rather than producing expressions that felt generic or disconnected, the model maintained the emotional continuity of the protagonist.
SO005 Google Play Kling AI: AI Image&Video Maker - Apps on Google Play Generate up to 15 seconds in native 1080p or cinema-grade 4K, then use Video Extension to create videos up to 3 minutes long.
SO006 Apple App Store Kling AI: AI Image&Video Maker App - App Store Standard plan (monthly) $10.00.
SO007 Kling AI Release Notes Kling 3.0 Model Now Fully Rolled Out, Setting a New Benchmark in AI Storytelling The new model integrates multiple tasks and supports longer video generation up to 15 seconds.
SO008 Kling AI Release Notes Kling IMAGE 3.0 Model officially launched The model supports direct 2K and 4K ultra-high-definition output.
SO009 Kuaishou Technology Investor Relations | Kuaishou Technology Kuaishou is a leading content community and social platform in China and globally.
SO010 Kuaishou Technology Annual & Interim Reports - Kuaishou Technology
SO011 HKEXnews ANNUAL REPORT 2025 In the fourth quarter of 2025, Kling AI achieved revenue of RMB340 million.
SO012 Kuaishou Technology Company News - Kuaishou Technology
SO013 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million in December 2025 Kling AI achieved monthly revenue exceeding USD20 million in December 2025, corresponding to an ARR of USD240 million.
SO014 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million in December 2025 (PDF) As of December 2025, Kling AI serves over 60 million creators worldwide.
SO015 Kuaishou Technology Kling AI Launches 3.0 Model, Ushering in an Era Where Everyone Can Be a Director Since its launch in June 2024, Kling AI now serves over 60 million creators worldwide.
SO016 Kuaishou Technology Kling AI Launches 3.0 Model, Ushering in an Era Where Everyone Can Be a Director (PDF) The model series features extended video duration of up to 15 seconds and native audio generation across multiple languages.
SO017 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results In the first quarter of 2026, Kling AI generated revenue of over RMB650 million.
SO018 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results (PDF) In March 2026, the annualized revenue run rate (ARR) of Kling AI was approximately USD500 million.
SO019 TechNode Kuaishou’s Kling AI raises nearly $3 billion in funding The financing will support Kling AI’s transition to independent commercial operations.
SO020 CNBC Kuaishou shares jump after Tencent joins $2.8 billion raise for Kling AI subsidiary Tencent was investing $200 million in the deal.
SO021 Business20 Channel Kling AI Raises $2.8B at $15B Valuation With Backing From Alibaba, Tencent and Baidu Kuaishou's ownership falls from 100% to about 68.33%, retaining a controlling stake.
SO022 QUASA Kling AI Raises $2.8 Billion at $18 Billion Valuation The transaction set a $15 billion pre-money valuation and $18 billion post-money valuation.
SO023 Kuaishou English Website Management Team
SO024 Cyberspace Administration of China Notice on Issuing the Measures for Labeling of AI-Generated Synthetic Content Service providers shall add conspicuous notification labels to video starting screens and metadata-based implicit labels.
SO025 China Law Translate Measures for Labeling of AI-Generated Synthetic Content App distribution platforms are required to verify labeling measures during app listing checks.
SO026 MLQ Kuaishou's Kling AI Raises $2.8B at $18B Valuation With Tencent, Alibaba Backing Kuaishou will retain a 68.33% controlling interest in Kling AI through subsidiary Beijing Kling after the transaction.
SM001 Kling AI Kling AI: Next-Gen AI Video & Image Generator All-New KlingAI 3.0 Series — All in One, One for All.
SM002 Kuaishou Technology Kling AI Monthly Revenue Exceeds USD20 Million, Annualized Revenue Run Rate Hits USD240 Million As of December 2025, Kling AI serves over 60 million creators worldwide, having generated more than 600 million videos and established partnerships with over 30,000 enterprise users.
SM003 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results In the first quarter of 2026, Kling AI generated revenue of over RMB650 million, representing year-over-year growth of more than 300%.
SM004 Kuaishou Technology 2025 Annual Report Kuaishou is a leading content community and social platform in China and globally.
SM005 Apple App Store KLING AI: The Next-Gen AI Creative Studio Generate up to 15 seconds in native 1080p or cinema-grade 4K, then use Video Extension to create videos up to 3 minutes long.
SM006 Google Play Kling AI Kling AI is an image and video generation and editing studio powered by AI.
SM007 MarketsandMarkets Generative AI Market by Offering, Data Modality, Application - Global Forecast to 2033 The global generative AI market is estimated at USD 185.45 billion in 2026 and is projected to reach USD 1,658.97 billion by 2033.
SM008 Statista Generative Artificial Intelligence - Worldwide The Generative Artificial Intelligence market is expected to see significant growth in the coming years, with a forecasted CAGR of over 24.4% from 2023 to 2030.
SM009 Adwave AI Video Statistics 2026: What the Data Says About Adoption, Cost, and Quality The AI video generator market was valued at roughly $788.5 million in 2025 and is projected to reach about $3.44 billion by 2033, a CAGR of 20.3%.
SM010 Coherent Market Insights AI Video Generation in 2026 Trends Reshaping the Industry The leaders in this space will not simply be the models that create the most eye-catching clip on first viewing. They will be the ones that offer repeatable control, flexible workflows, production-ready outputs, and practical ways to integrate generation into real creative systems.
SM011 Cyberspace Administration of China Measures for Labeling of AI-Generated Synthetic Content Service providers shall add conspicuous notification labels to video starting screens and metadata-based implicit labels.
SM012 China Law Translate Measures for Labeling of AI-Generated Synthetic Content App distribution platforms are required to verify labeling measures during app listing checks.
SM013 DigitalApplied OpenAI Just Shut Down Sora: What Comes Next for AI Video in 2026? The AI video generation sector is growing at a 34.2% compound annual growth rate.
SM014 Runway Runway A new production model for media.
SM015 Runway Pricing Explore the different plans for Runway.
SM016 Runway Introducing Gen-4 Consistent characters, locations and objects across scenes.
SM017 Google DeepMind Veo Create high-quality videos in a wide range of cinematic and visual styles.
SM018 Google DeepMind Veo 3 Veo 3 sets a new standard in video generation quality, excelling in physics, realism and prompt adherence.
SM019 Pika Pika Create videos from your ideas.
SM020 Pika Pricing Choose the plan that works for you.
SM021 Luma AI Dream Machine Create beautiful, realistic shots fast from text and images.
SM022 Luma AI Pricing Simple plans for creators and teams.
SM023 Atlas Cloud Kling vs Wan vs Seedream: Which AI Video Model Is Best in 2026? Kling — Known for cinematic motion and strong character consistency.
SM024 Business Wire Alibaba Cloud Unveils Wan2.1, the Industry's First Open-Source Video Foundation Model Suite with MoE Architecture Wan2.1 is the industry's first open-source video foundation model suite with MoE architecture.
SM025 Kling AI Blog Create E-commerce Product Videos in Native 4K with Kling AI Native 4K mode in Kling AI is built for advertising and commercial production where turnaround windows are tight, but quality is non-negotiable.
SM026 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results (PDF) In March 2026, the annualized revenue run rate (ARR) of Kling AI was approximately USD500 million.
SM027 Kuaishou Technology Kling AI Maintains Global Leadership in AI Video Generation, Launches 3.0 Model Series Since its launch in June 2024, Kling AI now serves over 60 million creators worldwide. To date, it has produced more than 600 million videos and forged partnerships with more than 30,000 enterprise clients.
SP001 Kling AI Kling AI: Next-Gen AI Video & Image Generator All-New KlingAI 3.0 Series — All in One, One for All.
SP002 Kuaishou Technology Kling AI Maintains Global Leadership in AI Video Generation, Launches 3.0 Model Series Since its launch in June 2024, Kling AI now serves over 60 million creators worldwide.
SP003 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million Kling AI serves over 60 million creators worldwide, having generated more than 600 million videos and established partnerships with over 30,000 enterprise users.
SP004 Runway Runway An all-in-one cloud-based creative platform that offers endless ways to generate and edit video, images and audio in one workspace.
SP005 Runway Pricing Explore the different plans for Runway.
SP006 Runway Introducing Gen-4 Consistent characters, locations and objects across scenes.
SP007 Google DeepMind Veo Create high-quality videos in a wide range of cinematic and visual styles.
SP008 Google DeepMind Veo 3 Veo 3 lets you add sound effects, ambient noise, and even dialogue to your creations — generating all audio natively.
SP009 Pika Pika Create videos from your ideas.
SP010 Pika Pricing Choose the plan that works for you.
SP011 Luma Pricing Creative agents that make you prolific.
SP012 Luma Dream Machine Luma unifies specialized multimodal models into one continuous workflow, advancing creative work from concept to final delivery.
SP013 Atlas Cloud Kling vs Wan vs Seedream AI Video Models 2026 Kling — Known for cinematic motion and strong character consistency.
SP014 AI Indigo AI Video Generators Compared: Sora, Runway, Kling & More in 2026 Google Veo 3.1 is arguably the best overall AI video generator in mid-2026.
SP015 Sean Kim AI Video Generation Comparison: Sora, Runway, Kling, and Pika Kling’s facial animation accuracy is particularly impressive — identity consistency in multi-shot sequences and expression preservation across fast cuts.
SP016 DualView Best AI Video Models Runway Gen-4.5 took the #1 spot immediately upon release.
SP017 WaveSpeed Kling AI API Pricing The real budget comes from model choice, seconds generated, resolution, audio, retries, review, and the share of outputs that are good enough to use.
SP018 CostBench Kling API Pricing Kling API offers usage-based pricing from $0.084–$0.420 per second of video as of August 2026.
SP019 OpenAI Help Center What to know about the Sora discontinuation Sora is being discontinued.
SP020 Business Wire Alibaba Cloud Unveils Wan2.1, the Industry's First Open-Source Video Foundation Model Suite with MoE Architecture Wan2.1 is the industry's first open-source video foundation model suite with MoE architecture.
SP021 DigitalApplied OpenAI Just Shut Down Sora: What Comes Next for AI Video in 2026? Runway, Kling, Google Veo, and Pika are building sustainable businesses around genuinely useful tools.
SP022 Kling AI Blog How Should Creators Think About Kling AI Credits? Standard — $6.99 / Month ... 660 Credits per month.
SP023 Kling AI Blog What Is Multi-Shot in Kling VIDEO 3.0? Shot limit: Up to 6 shots.
SP024 Kling AI Blog Kling 3.0 Credit Cost Guide The Kling VIDEO 3.0 series supports video generation up to 15 seconds.
SP025 Apple App Store KLING AI: The Next-Gen AI Creative Studio Pro plan (monthly) $37.00.
SI001 Kling AI Blog How Should Creators Think About Kling AI Credits? Standard — $6.99 / Month ... 660 Credits per month.
SI002 Apple App Store KLING AI: The Next-Gen AI Creative Studio Standard plan (monthly) $10.00 ... Pro plan (monthly) $37.00 ... Premier plan (monthly) $92.00.
SI003 CostBench Kling API Pricing Kling API offers usage-based pricing from $0.084–$0.420 per second of video as of August 2026.
SI004 WaveSpeed Kling AI API Pricing The real budget comes from model choice, seconds generated, resolution, audio, retries, review, and the share of outputs that are good enough to use.
SI005 KLING AI Terms of Paid Service Credit have an expiration date ... purchased Credit validity period is 2 years.
SI006 KLING AI Terms of Service Output may not always be accurate.
SI007 KLING AI Privacy Policy We store your Data on our servers located in Singapore.
SI008 Kling AI Blog Create E-commerce Product Videos in Native 4K with Kling AI A five-second clip at high quality usually takes about two minutes.
SI009 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million Kling AI achieved monthly revenue exceeding USD20 million in December 2025, corresponding to an ARR of USD240 million.
SI010 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results In the first quarter of 2026, Kling AI generated revenue of over RMB650 million.
SI011 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results (PDF) In March 2026, the annualized revenue run rate (ARR) of Kling AI was approximately USD500 million.
SI012 Kuaishou Technology Kling AI Maintains Global Leadership in AI Video Generation, Launches 3.0 Model Series Kling AI launched the Team Plan, supporting real-time collaborative creation for up to 15 members.
SI013 Kuaishou Technology 2025 Annual Report Total available funds reached RMB117.7 billion as of March 31, 2026.
SI014 TechNode Kuaishou’s Kling AI raises nearly $3 billion in funding The financing will support Kling AI’s transition to independent commercial operations.
SI015 CNBC Kuaishou shares jump after Tencent joins raise for Kling AI subsidiary Tencent was investing $200 million in the deal.
SI016 Pandaily Kuaishou Kling AI Spin-Off Hits $3 Billion Funding Kuaishou is spinning off Kling AI and bringing in outside capital.
SI017 Business20 Channel Kling AI Raises $2.8B at $15B Valuation With Backing From Alibaba, Tencent and Baidu Kuaishou's ownership falls from 100% to about 68.33%, retaining a controlling stake.
SI018 QUASA Kling AI Raises $2.8 Billion at $18 Billion Valuation The transaction set a $15 billion pre-money valuation and $18 billion post-money valuation.
SI019 MLQ Kuaishou's Kling AI Raises $2.8B at $18B Valuation With Tencent, Alibaba Backing Kuaishou will retain a 68.33% controlling interest in Kling AI.
SI020 Trustpilot Klingai is rated "Bad" with 1.3 / 5 on Trustpilot Subscription Refund Payment Cancellation Spam Service Customer service Customer communications Application Product
SI021 CostBench Kling API Pricing Overview Resource packages offer 10-30% bulk discounts.
SI022 WaveSpeed Kling API Cost Per Usable Video Cheap does not always mean cost-saving. Unusable generations are expensive.
SI023 KLING AI Paid service membership terms The validity period of Credit purchased is 2 years.
SI024 KLING AI Commercial-use terms in membership KLING AI members' use of the Output for commercial purposes is not restricted.
SI025 OpenAI Help Center What to know about the Sora discontinuation Sora is being discontinued.
SI026 Kling AI Pricing Pricing page
SE001 Kling AI API Kling 3.0 Model Now Fully Rolled Out, Setting a New Benchmark in AI Storytelling.
SE002 Kling AI About Us Kling AI
SE003 Kling AI Community Kling AI
SE004 Kling AI Resources ok
SE005 Kling AI VIDEO 3.0 Release Notes The new model integrates multiple tasks — such as Text-to-Video, Image-to-Video, Reference-to-Video, video content addition and removal, video modification and transformation — into a unified, native multimodal training model.
SE006 Kling AI IMAGE 3.0 Release Notes The model supports direct 2K and 4K ultra-high-definition output.
SE007 Kling AI Terms of Service Output may not always be accurate.
SE008 Kling AI Privacy Policy We store your Data on our servers located in Singapore.
SE009 Kuaishou Technology Kling AI Maintains Global Leadership in AI Video Generation, Launches 3.0 Model Series The model series features major upgrades in consistency, photorealistic output, extended video duration of up to 15 seconds, and native audio generation.
SE010 Kling AI Blog What Is Multi-Shot in Kling VIDEO 3.0? Shot Limit: Up to 6 shots.
SE011 Kling AI Blog Kling 3.0 Credit Cost Guide The Kling VIDEO 3.0 series supports video generation up to 15 seconds, with flexible duration from 3 to 15 seconds.
SE012 Kling AI Blog How Should Creators Think About Kling AI Credits? The 4K release introduced cinema-grade native 4K video output for the Kling VIDEO 3.0 series at 30 credits per second.
SE013 Kling AI Blog Create E-commerce Product Videos in Native 4K with Kling AI The system also includes native audio capabilities, producing sound effects and speech alongside the visuals for a truly immersive experience.
SE014 Kling AI Blog Beyond Realism: How Kling Powered L'Ultimo Uomo Reale Kling preserved that identity consistently across the production.
SE015 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results Kling AI launched the Team Plan, supporting real-time collaborative creation for up to 15 members.
SE016 Kling AI Release Notes IMAGE 3.0 Technical Roadmap The IMAGE O3 model introduces Visual Chain-of-Thought (vCoT) for the first time in the AI generation field.
SE017 Apple App Store KLING AI: The Next-Gen AI Creative Studio Generate up to 15 seconds in native 1080p or cinema-grade 4K, then use Video Extension to create videos up to 3 minutes long.
SE018 Google Play Kling AI: AI Image&Video Maker Kling AI is an image and video generation and editing studio powered by AI.
SE019 Cyberspace Administration of China Measures for Labeling of AI-Generated Synthetic Content Service providers shall add conspicuous notification labels to video starting screens and metadata-based implicit labels.
SE020 China Law Translate Measures for Labeling of AI-Generated Synthetic Content App distribution platforms are required to verify labeling measures during app listing checks.
SE021 Kling AI paid terms Terms of Paid Service The service fee charging scheme is subject to the announcement on the front-end interface of the relevant paid service product.
SE022 Sean Kim AI Video Generation Comparison: Sora, Runway, Kling, and Pika Kling’s facial animation accuracy is particularly impressive — identity consistency in multi-shot sequences and expression preservation across fast cuts.
SE023 DualView Best AI Video Models Kling O1 is the world's first unified multimodal video model, combining 18+ video tasks into a single platform.
SE024 Curious Refuge The Last Real Man | AI Short Film The campaign won Cannes Lions recognition and used Kling AI.
SE025 Intelligent CIO Global filmmakers leverage Kling AI to push the boundaries of storytelling at Cannes Global filmmakers leverage Kling AI to push the boundaries of storytelling at Cannes.
SU001 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million As of December 2025, Kling AI serves over 60 million creators worldwide, having generated more than 600 million videos and established partnerships with over 30,000 enterprise users.
SU002 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results Kling AI launched the Team Plan, supporting real-time collaborative creation for up to 15 members.
SU003 Kuaishou Technology Kling AI Launches 3.0 Model, Ushering in an Era Where Everyone Can Be a Director Since its launch in June 2024, Kling AI now serves over 60 million creators worldwide. To date, it has produced more than 600 million videos and forged partnerships with more than 30,000 enterprise clients.
SU004 Kuaishou Technology Kuaishou Technology Announces Fourth Quarter and Full Year 2025 Financial Results Kling AI innovations in foundational models and product features have paved the way for widespread commercial applications across professional creative sectors, including marketing, e-commerce, film and television, short plays, animation and gaming.
SU005 Apple App Store KLING AI: The Next-Gen AI Creative Studio 4.7 out of 5. 28K Ratings.
SU006 Google Play Kling AI: AI Image&Video Maker 537K reviews. 10M+ Downloads.
SU007 Kling AI Blog Beyond Realism: How Kling Powered L'Ultimo Uomo Reale Kling preserved that identity consistently across the production, enabling the creative team to move freely between locations, compositions and visual styles.
SU008 Kling AI Blog Kling AI Introduces Native 4K Video Model Since its launch, Kling 4K has already been adopted across a wide range of creative industries, from Hollywood production teams to independent creators, from animation studios to advertising agencies.
SU009 Kling AI Blog How to Create 4K E-commerce Product Videos with AI? Professional-quality product videos once demanded a full production team. Kling AI makes that process dramatically faster and more accessible.
SU010 Kling AI Release Notes Kling 3.0 Model Now Fully Rolled Out, Setting a New Benchmark in AI Storytelling The new model integrates multiple tasks — such as Text-to-Video, Image-to-Video, Reference-to-Video, video content addition and removal, video modification and transformation — into a unified, native multimodal training model.
SU011 Kling AI Release Notes Kling IMAGE 3.0 Model officially launched This provides robust support for professional work such as film storyboards, concept art, pre-visualization, and scene design.
SU012 Kling AI Blog Kling 3.0 15s Video: Master Narrative Control & Custom Duration The Kling VIDEO 3.0 series supports video generation up to 15 seconds, with flexible duration from 3 to 15 seconds.
SU013 Kling AI Blog Kling VIDEO 3.0 Multi-Shot: Create Structured Cinematic Sequences Shot Limit: Up to 6 shots.
SU014 Variety How 'House of David' Used AI for a Goliath Origin Story Sequence In all, season one incorporates 72 shots that involved use of AI.
SU015 Variety AI Disruption Dominates Filmart According to Chen Yi, founder of Swords VFX provider Timeaxis Studios, Kling AI was integrated into every stage of the pipeline, from rapidly generated pre-vis material to generating final effects plates for compositing.
SU016 Little Black Book The RealReal: The Last Real Man The RealReal has released a film that examines the shifting line between what’s real and what isn’t.
SU017 Curious Refuge The Last Real Man | AI Short Film This Cannes Lions 2026 Silver and Bronze Lion-winning AI film directed by Sebastian Strasser and produced by Lipstick for Team One using Kling AI.
SU018 Intelligent CIO Global filmmakers leverage Kling AI to push the boundaries of storytelling at Cannes With House of David, Kling AI served as its core foundation model and benchmark tool.
SU019 Yahoo Finance Global filmmakers leveraging Kling AI Across both season one and two, Kling AI has generated the vast majority of its production shots, leading the share of its AI workflows for the show.
SU020 Selfstorming The RealReal: L'Ultimo Uomo Reale The film achieved nearly one million views shortly after release, sparking widespread industry debate regarding the ethics of AI in advertising.
SU021 Trustpilot Klingai is rated Bad with 1.3 / 5 on Trustpilot Klingai is rated "Bad" with 1.3 / 5 on Trustpilot.
SU022 Kling AI paid terms Terms of Paid Service Once the Paid Service (including any Membership Service and Credit service) is activated, the fee paid for the Paid Service is non-refundable.
SU023 Kling AI Service Agreement The Services are offered and provided to you on an "as is" and "as available" basis at your sole risk.
SU024 Kling AI Privacy Policy We store and process the content and Data you upload, generate, or access through the Platform.
SU025 Kling AI Blog AI image and video generation tutorials, practical tips, and product updates AI image and video generation tutorials, practical tips, and product updates.
SR001 Cyberspace Administration of China Measures for Labeling of AI-Generated Synthetic Content Service providers providing generated synthetic-content download, copy, or export functions shall ensure the files contain compliant explicit labels.
SR002 China Law Translate Measures for Labeling of AI-Generated Synthetic Content Internet application distribution platforms shall check materials related to the labeling of generated synthetic content.
SR003 Deep Lex China AI Regulation Tracker As of 28 February 2026, 796 generative AI services and 481 applications or functions have completed registration.
SR004 Regulations.AI China Summary China is rapidly developing a comprehensive AI regulatory framework, balancing innovation with stringent controls on content, data security, and national interests.
SR005 Regulations.AI Measures for the Identification of AI-Generated Synthetic Content App distribution platforms are required to verify labeling measures during app listing checks.
SR006 Securiti China AI Regulatory Landscape Organizations are encouraged to establish robust governance structures and incident response teams and mechanisms.
SR007 Cimplifi The AI Regulation Landscape for 2026 Organizations developing, deploying, or relying on AI systems operating across borders will need to navigate overlapping regimes that reflect very different regulatory philosophies.
SR008 Kling AI Service Agreement The Services are offered and provided to you on an "as is" and "as available" basis at your sole risk.
SR009 Kling AI Privacy Policy We store and process the content and Data you upload, generate, or access through the Platform.
SR010 Kling AI paid terms Terms of Paid Service Once the Paid Service (including any Membership Service and Credit service) is activated, the fee paid for the Paid Service is non-refundable.
SR011 Trustpilot Klingai is rated Bad with 1.3 / 5 on Trustpilot Klingai is rated "Bad" with 1.3 / 5 on Trustpilot.
SR012 Google Play Kling AI: AI Image&Video Maker 10M+ Downloads.
SR013 Apple App Store KLING AI: The Next-Gen AI Creative Studio 4.7 out of 5. 28K Ratings.
SR014 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results Kling AI generated revenue of over RMB650 million in the first quarter, representing year-over-year growth of more than 300.0%.
SR015 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million Kling AI serves over 60 million creators worldwide, having generated more than 600 million videos and established partnerships with over 30,000 enterprise users.
SR016 Kuaishou Technology Kuaishou Technology Announces Fourth Quarter and Full Year 2025 Financial Results Kling AI innovations in foundational models and product features have paved the way for widespread commercial applications across professional creative sectors.
SR017 Kuaishou Technology 2025 Annual Report PDF We remain committed to investing in AI technology and have made remarkable progress. Our multimodal large video generation model, Kling AI, accelerated its iteration throughout 2025.
SR018 Variety AI Disruption Dominates Filmart Kling AI was integrated into every stage of the pipeline, from rapidly generated pre-vis material to generating final effects plates for compositing.
SR019 Variety How House of David Used AI for a Goliath Origin Story Sequence The sequence might have otherwise taken four or five months in a traditional process.
SR020 Intelligent CIO Global filmmakers leverage Kling AI to push the boundaries of storytelling at Cannes With House of David, Kling AI served as its core foundation model and benchmark tool.
SR021 Yahoo Finance Global filmmakers leveraging Kling AI Across both season one and two, Kling AI has generated the vast majority of its production shots.
SR022 TechNode Kuaishou’s Kling AI raises nearly $3 billion in funding The financing will support Kling AI’s transition to independent commercial operations.
SR023 Quasa Kling AI raises $2.8 billion at $18 billion valuation The deal dilutes Kuaishou's stake to about 68% ahead of a planned Hong Kong listing.
SR024 CNBC Kuaishou shares jump after Tencent joins $2.8 billion raise for Kling AI subsidiary Kuaishou shares jump after Tencent joins $2.8 billion raise for Kling AI subsidiary.
SR025 Economic Times CIO / Reuters US licenses Nvidia to export chips to China, official says The commerce department has started issuing licenses to Nvidia to export its H20 chips to China.
SR026 Bureau of Industry and Security Department of Commerce Revises License Review Policy for Semiconductors Exported to China BIS will now review export license applications for the Nvidia H200, AMD MI325X, and similar chips on a case-by-case basis provided certain security requirements are met.
SR027 Bureau of Industry and Security Department of Commerce Revises License Review Policy for Semiconductors Exported to China PDF The rule is effective immediately upon publication in the Federal Register.
SR028 Kling AI Blog Kling AI Introduces Native 4K Video Model Since its launch, Kling 4K has already been adopted across a wide range of creative industries.
SR029 Kling AI Release Notes Kling 3.0 Model Now Fully Rolled Out, Setting a New Benchmark in AI Storytelling The new model integrates multiple tasks into a unified, native multimodal training model.
SR030 Kling AI Blog AI image and video generation tutorials, practical tips, and product updates AI image and video generation tutorials, practical tips, and product updates.
SV001 CNBC Kuaishou shares jump after Tencent joins $2.8 billion raise for Kling AI subsidiary Kuaishou shares jump after Tencent joins $2.8 billion raise for Kling AI subsidiary.
SV002 TechNode Kuaishou’s Kling AI raises nearly $3 billion in funding The financing will support Kling AI’s transition to independent commercial operations.
SV003 Quasa Kling AI raises $2.8 billion at $18 billion valuation The transaction set a $15 billion pre-money valuation and $18 billion post-money valuation.
SV004 Kuaishou Technology Kuaishou Technology Announces First Quarter 2026 Unaudited Financial Results In March 2026, the annualized revenue run rate (ARR) of Kling AI was approximately USD500 million.
SV005 Kuaishou Technology Kling AI Annualized Revenue Run Rate Hits USD240 Million Kling AI serves over 60 million creators worldwide, having generated more than 600 million videos and established partnerships with over 30,000 enterprise users.
SV006 Kuaishou Technology Kuaishou Technology Announces Fourth Quarter and Full Year 2025 Financial Results Kling AI innovations in foundational models and product features have paved the way for widespread commercial applications across professional creative sectors.
SV007 Kuaishou Technology 2025 Annual Report PDF We remain committed to investing in AI technology and have made remarkable progress. Our multimodal large video generation model, Kling AI, accelerated its iteration throughout 2025.
SV008 Variety AI Disruption Dominates Filmart Kling AI was integrated into every stage of the pipeline, from rapidly generated pre-vis material to generating final effects plates for compositing.
SV009 Variety How House of David Used AI for a Goliath Origin Story Sequence The sequence might have otherwise taken four or five months in a traditional process.
SV010 Intelligent CIO Global filmmakers leverage Kling AI to push the boundaries of storytelling at Cannes With House of David, Kling AI served as its core foundation model and benchmark tool.
SV011 Yahoo Finance Global filmmakers leveraging Kling AI Across both season one and two, Kling AI has generated the vast majority of its production shots.
SV012 Google Play Kling AI: AI Image&Video Maker 10M+ Downloads.
SV013 Apple App Store KLING AI: The Next-Gen AI Creative Studio 4.7 out of 5. 28K Ratings.
SV014 Trustpilot Klingai is rated Bad with 1.3 / 5 on Trustpilot Klingai is rated "Bad" with 1.3 / 5 on Trustpilot.
SV015 Kling AI Service Agreement The Services are offered and provided to you on an "as is" and "as available" basis at your sole risk.
SV016 Kling AI paid terms Terms of Paid Service Once the Paid Service (including any Membership Service and Credit service) is activated, the fee paid for the Paid Service is non-refundable.
SV017 Kling AI Privacy Policy We store and process the content and Data you upload, generate, or access through the Platform.
SV018 Cyberspace Administration of China Measures for Labeling of AI-Generated Synthetic Content Service providers providing generated synthetic-content download, copy, or export functions shall ensure the files contain compliant explicit labels.
SV019 CompaniesMarketCap Adobe market cap Market cap: $104.94 Billion USD
SV020 CompaniesMarketCap Adobe revenue Revenue in 2026 (TTM): $25.19 Billion USD
SV021 CompaniesMarketCap Duolingo market cap Market cap: $6.21 Billion USD
SV022 CompaniesMarketCap Duolingo revenue Revenue in 2026 (TTM): $1.09 Billion USD
SV023 CompaniesMarketCap GitLab market cap Market cap: $7.19 Billion USD
SV024 CompaniesMarketCap GitLab revenue Revenue in 2026 (TTM): $1.00 Billion USD
SV025 CompaniesMarketCap C3.ai market cap Market cap: $1.54 Billion USD
SV026 CompaniesMarketCap C3.ai revenue Revenue in 2026 (TTM): $0.25 Billion USD
SV027 Kling AI Blog Kling AI Introduces Native 4K Video Model Since its launch, Kling 4K has already been adopted across a wide range of creative industries.
SV028 Kling AI Release Notes Kling 3.0 Model Now Fully Rolled Out, Setting a New Benchmark in AI Storytelling The new model integrates multiple tasks into a unified, native multimodal training model.
SV029 Regulations.AI China Summary China is rapidly developing a comprehensive AI regulatory framework, balancing innovation with stringent controls on content, data security, and national interests.
SV030 Bureau of Industry and Security Department of Commerce Revises License Review Policy for Semiconductors Exported to China BIS will now review export license applications for the Nvidia H200, AMD MI325X, and similar chips on a case-by-case basis provided certain security requirements are met.