如果你收到了邀请码,请在下方输入
Enter your invite code to continue
电影,另一种形式。 Cinema, in another form.
麦肯锡眼里,三场结构性的转变。Three shifts McKinsey now considers structural.
美国电视与电影占视频观看时间的份额:2019 年 61% → 2025 年 50%。线性电视年降 4%,社交视频年增 14%。美国成人日均视频消费约 7 小时。TV and film: 61% → 50% of US video viewing share since 2019. Linear TV declining –4% CAGR; social video +14%. Average US adult: ~7 hrs of video daily.
来源:麦肯锡《The attention equation》(2025年6月)Source: McKinsey, "The attention equation" (Jun 2025)
2024 年全球内容支出 1810 亿美元(不含体育)。美国约 1010 亿美元,预计每年下滑 ~2%,平台资金转向体育版权与许可内容。Global content spend reached $181B in 2024 (excl. sports). US spend (~$101B) is projected to decline ~2% per year as buyers shift to sports and licensed programming.
来源:McKinsey Content Spend Forecast (2025)Source: McKinsey Content Spend Forecast (2025)
麦肯锡测算三种可能结局:工作流加速、内容民主化、全新内容形态。历史经验显示,戏剧→电影、线性→流媒体、长→短 三次结构转变,旧赢家在 5 年内损失约 35% 收入。McKinsey models three potential outcomes: workflow scaling, content democratization, NEW FORMATS. Past disruptions (stage→cinema, linear→streaming, long→short) each contracted incumbent revenue ~35% within 5 years.
来源:McKinsey《What AI could mean for film and TV》(2026年1月)Source: McKinsey, "What AI could mean for film and TV" (Jan 2026)
世界不在等下一个伟大的故事。它在等那个建造下一种形式的人。The world isn't waiting for the next great story. It's waiting for whoever builds the next great form.
A24 用观众能感受的故事在票房上跑赢传统大片厂。A24 outpaces studios at the box office with stories audiences feel.
《过往人生》 1200 万美元成本拿下 4270 万美元票房。《晒后假日》《坠落的审判》皆是证明。Past Lives earns $43M on a $12M budget. Aftersun, Anatomy of a Fall, all proof.
大部分短视频是垃圾。品味市场是开阔的。Most short-form video is slop. The taste market is wide open.
品质,是这个时代最反共识的押注。我们为算法做的,就是它。Quality is the new contrarian bet. We're building it for the algorithm.
大多数 AI 视频工具,优化的是「看起来合理」,不是「让人记住」。生成出来的那一刻,就被忘了。Most AI video tools optimize for 'plausible,' not for 'unforgettable.' The output is forgettable on arrival.
生成模型不懂什么样的故事能打动人。它们能渲染画面,渲染不了情感。Generation models don't know what makes a story land. They render frames, not feeling.
量产工厂用低质内容把信息流灌满。供给端,正在自己塌进自己里。Volume-first studios flood feeds with low-effort content. The supply curve is collapsing on itself.
观看时长在掉,转发率在缩。平台已经悄悄承认了这件事。Watch time is fracturing. Shares are drying up. The platforms quietly admit it.
画质、人才、分发,在同一年汇合。窗口已经打开。Quality, talent, and distribution converged in the same year. The window is open.
Sora 2、Veo 3、可灵 2.0、Seedance 1.0 已经能生成 1080p、人物稳定的电影级片段。Netflix 在 2025 Q2 财报里披露:原创剧集第一次用上 AI 生成画面。Sora 2, Veo 3, Kling 2.0, and Seedance 1.0 now produce 1080p clips with stable character consistency. Netflix shipped its first AI-generated final footage in Q2 2025.
《哪吒 2》全球票房 21.5 亿美元,制作成本 8000 万——史上最卖座的动画电影,背后是 138 家中国动画公司、4000+ 位艺术家。Ne Zha 2 grossed $2.15B worldwide on an $80M budget — highest-grossing animated film ever, made by 138 Chinese studios and 4,000+ artists.
2025 年全球短视频广告市场:1110 亿美元,预计 2028 年达 1460 亿美元(年复合增长率 13.7%)。来源:Statista。Global short-form ad market in 2025. Projected $146B by 2028 at 13.7% CAGR. Source: Statista.
钩子 · 张力 · 回报。A24 之所以是 A24 的三幕剧脊柱 — 现在压进 60 秒。Hook · tension · payoff. The same three-act spine that made A24 great — now in 60 seconds.
观众一直奖赏的是情感,不是工艺。AI 抬高了下限;品味抬高了上限。Audiences have always rewarded feeling over polish. AI raised the floor; taste raised the ceiling.
留存率,是新的票房。每一条片子都在第一时间告诉我们:什么故事打动了观众。Retention is the new box office. Every post tells us, instantly, what stories landed.
90 秒的电影,AI 做的,分量却是 90 分钟。 A 90-second film, made by AI, that lands like a 90-minute one.
Talent 在一端。Audience 在另一端。我们让两端相遇。Talent on one side. Audience on the other. We make them meet.
《哪吒 2》全球票房 21.5 亿美元(2025),史上最卖座动画电影,138 家中国工作室协作。《黑神话:悟空》销量 2500 万+ 份。中国 craft 已经全球化,但 IP 大多停留在国内。Ne Zha 2 grossed $2.15B worldwide on $80M (2025) — highest-grossing animated film ever. Black Myth: Wukong sold 25M+ units. Chinese craft is global; the IP stays trapped at home.
中国动画师收入是好莱坞的 1/5。加上 AI 工具,每分钟成本降到原来的 1/20。Chinese animators earn 1/5 of Hollywood. With AI tools, per-minute cost falls to ~1/20 of the original.
ReelShort、TikTok、小红书已经验证了竖屏 + 连载 + 算法分发。我们站在巨人肩上。ReelShort, TikTok, and Xiaohongshu proved vertical + episodic + algorithmic distribution. We stand on giants' shoulders.
一个 AI 原生平台,连接世界级新兴电影人才与渴望好故事的观众。 An AI-native platform connecting world-class emerging cinema talent with audiences hungry for great stories.
三层集成 — 输入审美,输出表现,每一帧都是训练信号。 Three integrated layers — taste in, performance out, every frame a training signal.
A 15-episode short cinema, not a one-off clip.A 15-episode short cinema, not a one-off clip.
15 × 90 秒 = 约 30 分钟 = 一部短片的体量,一集竖屏的呼吸。 15 × 90 sec = ~30 min = the scope of a short film; the breath of a vertical episode.
三幕剧 → 15 个钩子Three-act spine → 15 hooks. 每集独立成篇,合起来是一条弧。Each episode stands alone. Together they make an arc.
短剧的产品逻辑,Short drama's product logic, Oscar-short craft. 可以拆成集,可以剪成片。Splits into episodes, cuts into a film.
Episode 1 钩住,15 集 LTVHook in EP 1. 15 episodes of LTV. 比 single video 高 ~8x。~8× the LTV of a single video.
不只是数学。是勇气、好奇心、共情。孩子要的是故事,父母愿意为质感买单。Not just math. Courage, curiosity, empathy. Kids want stories. Parents pay for quality.
家庭、事业、爱情。期望、义务、独立。沉默的精疲力尽。Family, career, love. Expectation, obligation, independence. The quiet burnout.
爱情、梦想、家庭、归属。普世主题,不分国界 — 故事是 therapy,是 belonging。Love, dreams, family, belonging. Universal themes, no borders — story is therapy, connection, belonging.
竖屏电影。A24 的审美,为信息流而生。Vertical cinema. A24 sensibility, made for the feed.
她回到家。她想起了为什么离开。She comes home. She remembers why she left.
三个时区,两种语言,一场分手。Three time zones, two languages, one breakup.
她拿到了大办公室。她仍然不敢告诉母亲。She got the corner office. She still can't tell her mother.
有些东西无法翻译。她还是试着说。Some things don't translate. She tries anyway.
她回家一个星期。
她想起了为什么离开。
She comes home for one week.
She remembers why she left.
预期:200 万 + 播放 · 8%+ 转发率 · 第一天就建立品类权威。Forecast: 2M+ views · 8%+ share rate · category authority on day one.
三集试映 · 即将上线。 Three-episode preview · Streaming soon.
六个一句话故事。每个 60 秒电影。电影节质感,为算法准备。Six loglines. Each a 60-second film. Festival sensibility, algorithm-ready.
她回家一个星期。又想起了为什么离开。She comes home for one week and remembers why she left.
她搬进他的公寓。前任的画还挂在墙上。她没取下来。She moves into his apartment. His ex's paintings are still up. She leaves them.
她拿到了大办公室。她仍然不敢告诉母亲。She got the corner office. She still can't tell her mother how she feels.
三十,单身,在场唯一没哭的人。Thirty, single, and the only one not crying.
有些东西无法翻译。她还是试着说。Some things don't translate. She tries anyway.
给讨厌数学的小孩讲数学。A math story for kids who hate math.
2025 年全球短视频视频广告。预计 2028 年达 1460 亿美元(年复合增长率 13.7%)。来源:Statista。Global short-form video advertising in 2025. Projected $146B by 2028 at 13.7% CAGR. Source: Statista.
两年内 +40%。Thrive Capital 领投 — 也是 OpenAI 核心投资人。同一拨钱,押电影品味也押 AI。+40% in 2 years. Round led by Thrive Capital — also OpenAI's top investor. The same money pricing cinema taste is pricing AI.
麦肯锡测算:AI 大规模应用后的 5 年里,约 600 亿美元会被重新分配。我们正为这场重分配建公司。McKinsey models ~$60B as redistributable within 5 years of mass AI adoption. We're building for that redistribution.
电视与电影的份额从 2019 年 61% 跌至 2025 年 50%。来源:麦肯锡《The attention equation》。TV/film share of US viewing fell from 61% (2019) to 50% (2025). Source: McKinsey, "The attention equation."
来源:Statista (2025);Variety, Hollywood Reporter (2024年6月);McKinsey Content Spend Forecast (2025);McKinsey《The attention equation》(2025年6月)。 Sources: Statista (2025); Variety, Hollywood Reporter (Jun 2024); McKinsey Content Spend Forecast (2025); McKinsey, "The attention equation" (Jun 2025).
AI 模型谁都能调用。但几乎没人知道,哪种故事能在信息流里把人拽下来。Anyone can call an AI model. Almost no one knows what story will make a thumb stop scrolling.
一个由 A24 风格训练的叙事弧、钩子与情感触发器语料库。其他人没有的品味数据。An A24-trained corpus of arcs, hooks, and emotional triggers. The taste data nobody else has.
大多数创作者为工艺设计。我们为观看时长设计。不同的目标函数,不同的产出。Most creators design for craft. We design for watch time. Different objective, different output.
每一条视频都让模型更锋利。每一位观众都是一票。我们跑得越久,越难被复制。Every video sharpens the model. Every viewer is a vote. The longer we run, the harder we are to copy.
更多员工 → 更多内容。More headcount → more content.
线性。脆弱。上限是人。Linear. Fragile. Capped by humans.
更多数据 → 更好内容 → 更多观看。More data → better content → more views.
非线性。复利。上限是物理。Non-linear. Compounding. Capped by physics.
在 TikTok、Reels、Shorts 上发布。两个切入点共发 50-100 条 AI 生成短片,找到可复制的赢家。Launch on TikTok, Reels, Shorts. Ship 50-100 AI-generated short films across both wedges to find repeatable winners.
聚类有效内容:钩子、弧线、受众。把产能集中到前 20% 的格式上。Cluster what worked: hooks, arcs, audience segments. Concentrate production on the top quintile of formats.
围绕已验证的 IP 建频道。衍生角色、系列、受众垂直。Build channels around proven IP. Spin off characters, series, and audience-specific verticals.
聚合新兴电影人才 + AI 降本。制作首批 IP,与人才分账(类似 TikTok 创作者代理)。Aggregate emerging talents + AI cost reduction. Produce first IP, revenue share with talents (like a TikTok creator agency).
原生品牌故事 + 付费订阅。情感内容 CPM 是娱乐内容的 5 倍。Native sponsored stories + premium subscriptions. Emotional content CPMs run 5x entertainment.
积累自有内容版权。跨文化翻译:中国故事去美国,美国故事来亚洲。Accumulate proprietary content rights. Cross-cultural translation: Chinese stories to the US, US stories to Asia.
在 IP 语料上训练自有审美算法。邀请作者导演(Celine Song、Lulu Wang)。生态护城河。Train proprietary taste algorithm on IP corpus. Invite auteurs (Celine Song, Lulu Wang). Ecosystem moat.
观众投票决定哪部短片值得做成长片。算法是新的 development office。 Audiences vote on which shorts deserve to be features. The algorithm is the new development office.
懂艺术电影的不懂产品逻辑。懂产品的不懂叙事弧。懂中美桥梁的不懂内容。这三件,我们都懂。 Cinema people don't get product. Product people don't get arcs. Bridge people don't get content. We get all three.
两位接受电影 / 编剧训练的创始人。Wharton 电影研究、Wesleyan 电影、Columbia MFA 制片课程。Two filmmakers / screenwriters by training. Wharton Cinema Studies, Wesleyan Film, Columbia MFA producing.
BCG 战略 + TikTok 平台 + 投行 + Wharton 金融。竖屏叙事、留存、病毒传播、资本市场。BCG strategy + TikTok platform + investment banking + Wharton finance. Vertical storytelling, retention, virality, capital markets.
懂中国人才生态,也懂美国 / 全球观众。翻译跨文化影响力。Knows Chinese talent ecosystem AND US / global audience. Translates cross-cultural impact.
我们不是站在外面看这个行业——我们就在里面。 We're not observing this industry from outside. We're in it.
懂艺术电影的不懂产品逻辑。懂产品的不懂叙事弧。懂中美桥梁的不懂内容。这三件,我们都懂。 Cinema people don't get product. Product people don't get arcs. Bridge people don't get content. We get all three.
电影人 · 编剧 · TikTok 产品战略 · BCG · UPenn Wharton + Columbia MBAFilmmaker · Screenwriter · TikTok Product Strategy · BCG · UPenn Wharton + Columbia MBA
电影创作者 · 编剧 · Stanford GSB · Wesleyan · 前投行Filmmaker · Screenwriter · Stanford GSB · Wesleyan · ex-Investment Banker
短片,是故事证明自己值得做长的舞台。 Short-form is where stories prove they deserve to be longer.
麦肯锡指出:过去三次形态切换 — 戏剧→电影、线性→流媒体、长→短 — 每次让旧赢家在 5 年内损失约 35% 收入。我们用数据先测出哪个故事值得规模化。McKinsey: prior format shifts (stage→cinema, linear→streaming, long→short) cut incumbent revenue ~35% in 5 years. We test what's worth scaling — first.
每一集都在测试观众反应。表现最佳的角色、世界、弧线,成为长片的种子。而粉丝,从第一集起就跟着。Each episode tests audience response. Best-performing characters, worlds, and arcs become feature seeds. Fans were there from episode one.
新兴人才通过短片证明能力,赢得长片预算。我们是 Sundance + Y Combinator 的合体。Emerging talent proves capability through shorts, earns feature budgets. We are Sundance + Y Combinator combined.
从 90 秒起步。最终,会更大。 It will start as 90 seconds. It will end up bigger than that.
无论你是创作人才、投资人,还是相信 quality storytelling 的观众Whether you're a creative talent, an investor, or a believer in quality storytelling