AI News AI资讯 7h ago Updated 1h ago 更新于 1小时前 50

Kling AI Two Anniversary: Global Users Surpass 100 Million, Enterprise Customers Nearly 50,000 可灵AI两周年:全球用户突破1亿,企业客户近5万家

Reaching 100 million users in two years with quarterly revenue exceeding 650 million yuan, Kling AI's report card reads less like steady growth and more like a "violent acceleration" of the entire industry's pace. While most are still debating the "usability" of AI video generation, Kuaishou has used these figures to accomplish two things: first, proving that the commercialization pathway for AI video has been fully opened; and second, raising the survival threshold for latecomers to a suffocati 两年用户破亿,单季营收超6.5亿——可灵AI的这份成绩单,与其说是稳步增长,不如说是一场对整个行业节奏的“暴力提速”。当大多数人还在讨论AI视频生成的“可用性”时,快手用这串数字完成了两件事:一是证明了AI视频的商业化通路已彻底打通;二是把后来者的生存门槛,踩到了一个令人窒息的高度。

75
Hot 热度
70
Quality 质量
70
Impact 影响力

Analysis 深度分析

Reaching 100 million users in two years with quarterly revenue exceeding 650 million yuan, Kling AI's report card reads less like steady growth and more like a "violent acceleration" of the entire industry's pace. While most are still debating the "usability" of AI video generation, Kuaishou has used these figures to accomplish two things: first, proving that the commercialization pathway for AI video has been fully opened; and second, raising the survival threshold for latecomers to a suffocating height.

Data doesn’t lie, but it tells a story. The user base and enterprise clients both surged by 67%—a strikingly consistent ratio. This points to a stark reality: Kling AI’s growth is no longer fueled solely by casual consumer experimentation but driven by tangible business procurement from B2B clients. ARR (Annual Recurring Revenue) approached $500 million within a year, marking nearly 400% growth—a pace that ranks at the top of any SaaS sector. This indicates that for industries like advertising, film, e-commerce, and design, AI video generation has moved from being an "optional" add-on to a "must-have" item on the productivity procurement list. This is no longer conceptual; it is the real-world evolution of tools powered by actual investment.

The speed of technological iteration is equally staggering. From version 1.0 to 3.0 Omni, there have been 26 updates, 122 papers published, and 21 open-source projects launched in just one year. Behind these numbers is a near-"anxious" saturation attack on technology. Kuaishou clearly understands that on the track of AI video, the window of opportunity created by technological gaps is extremely narrow. It must prove its position at the forefront at the fastest possible frequency, build an ecological moat through papers and open-source contributions, and transform the evolution of model capabilities from a "release date" event into an "ongoing" process. This strategy is a dimensionality-reducing blow to competitors iterating at an order of magnitude slower pace.

However, beneath the spotlight, undercurrents are equally turbulent. How much of the revenue surge comes from "trial-based" purchases, and how much can be converted into long-term, highly sticky customer retention? As the technical barriers of generative video are gradually leveled, will price wars and feature homogenization immediately follow? Kling AI’s dominant lead is forcing the entire sector to accelerate its shakeout. Teams that possess only a single video generation function—lacking vast traffic scenarios and engineering deployment capabilities—will likely quickly become cannon fodder in this arms race.

The deeper competition may no longer be about specifications on a technical datasheet. At its core, Kling AI is built upon Kuaishou’s massive content ecosystem and years of accumulated expertise in video understanding. Its most formidable asset is not the model itself, but the seamless integration of the model with short video, live streaming, and e-commerce scenarios. This closed loop of "scenario-model-application" is a dimensionality-reducing advantage that no pure tech company can easily replicate. While other players are still wrestling with generating a few seconds of photorealistic footage, Kling AI may already be thinking about how to generate product-promotion short videos with one click or automatically edit highlight clips from livestreams.

Thus, Kling AI’s meteoric rise serves more as a signal: the "tech flexing" phase of AI video generation is drawing to a close, and the ruthless "commercial blade-to-blade" phase has begun. Revenue figures are both a moat and a target. For latecomers to challenge, what’s needed is not a better model, but an entirely different and more robust ecological niche. Meanwhile, the pioneer faces the task of avoiding being shackled by "growth" itself under its absolute lead—truly delving into the depths of industry to complete the thrilling leap from "tool" to "infrastructure." The race for the future has only just bared its fangs.

两年用户破亿,单季营收超6.5亿——可灵AI的这份成绩单,与其说是稳步增长,不如说是一场对整个行业节奏的“暴力提速”。当大多数人还在讨论AI视频生成的“可用性”时,快手用这串数字完成了两件事:一是证明了AI视频的商业化通路已彻底打通;二是把后来者的生存门槛,踩到了一个令人窒息的高度。

数据不会说谎,但数据会讲故事。用户数和企业客户数同步暴增67%,这个比例惊人地一致。这指向一个冰冷的现实:可灵AI的增长不再是单纯依靠C端用户的热情尝鲜,而是B端实实在在的商业采购在驱动。ARR(年化经常性收入)一年内逼近5亿美元,增长近400%,这种速度放在任何SaaS赛道都是顶尖水平。它意味着,广告、影视、电商、设计等行业的“生产力采购清单”上,AI视频生成已从“可选”跃升为“必选项”。这不再是概念,这是真金白银的生产工具迭代。

技术迭代的速度同样惊人。从1.0到3.0 Omni,一年26次更新、122篇论文、21个开源项目。这组数字背后是一种近乎“焦虑式”的技术饱和攻击。快手显然很清楚,在AI视频这条赛道上,技术代差带来的窗口期极短。它必须用最快的频率证明自己处于前沿,用论文和开源构筑生态护城河,把模型能力的进化从“发布日”变成“进行时”。这种打法,对那些迭代速度慢一个量级的竞品而言,是降维打击。

然而,光环之下,暗流同样汹涌。营收的猛增,有多少来自“尝鲜式”的采购,有多少能转化为长期、高粘性的客户留存?当生成式视频的技术壁垒被逐渐摊平,价格战和功能内卷是否立刻会到来?可灵AI的强势领跑,正在倒逼整个赛道加速洗牌。那些仅有单一视频生成功能、缺乏庞大流量场景和工程化落地能力的团队,恐怕会迅速沦为这场军备竞赛中的炮灰。

更深层的竞争,或许已不在技术参数表上。可灵AI的底层,是快手庞大的内容生态和对视频理解的多年积累。它最可怕的不是模型本身,而是模型与短视频、直播、电商场景的无缝咬合。这种“场景-模型-应用”的闭环,是任何纯技术公司都难以复制的降维优势。其他玩家还在纠结于生成几秒钟的逼真画面时,可灵AI可能已经在思考如何一键生成带货短视频,或者自动剪辑直播高光片段了。

所以,可灵AI的狂飙,更像是一个信号:AI视频生成的“技术秀肌肉”阶段正在落幕,残酷的“商业拼刺刀”阶段已然开场。营收数字是护城河,也是靶子。后来者若想挑战,需要的不是更好的模型,而是完全不同且更坚固的生态位。而领航者面临的,是如何在绝对领先的增速下,避免被“增长”本身绑架,真正扎进产业深处,去完成从“工具”到“基础设施”的惊险一跃。这场关于未来的竞赛,才刚刚露出它的獠牙。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

视频生成 视频生成 多模态 多模态 大模型 大模型
Share: 分享到: