WAKENEX觉醒联盟
CASES · LEVERAGE · RULES案例 · 杠杆 · 规则

One-Person Company一人公司研究

Who built what alone — and what the numbers actually prove.

一个人做成了什么，那些数字又到底证明了什么。

A working bench for people building with AI and leverage. Every case shows its source and whether the revenue is verified or only self-reported; next to it sit the base rates, platform rules and tax changes that decide how far one person can go. No income screenshots, no promises.

给想用 AI 和杠杆做一人公司的人准备的一张实验台。每个案例都标明出处，并说明收入是“已核实”还是“仅自述”；旁边放着基准数据、平台规则和税务变化，这些才决定一个人能走多远。不晒收入截图，不承诺结果。

Kept this week本周留下
3
Sources来源
20
WAKENEX piecesWAKENEX 内容
8

From the WAKENEX solo labWAKENEX 一人公司实验室

Our own essays and notes on this topic: thesis, mechanism, the strongest case against us.

WAKENEX 在本主题的长文与短评：论点、机制、最强反方。

Systems Research系统研究

Going solo with AI deletes your last independent check.一个人用 AI 单干，等于删掉了你最后一道独立的纠错。

Bloomberg ran a how-to on October 25 for running an AI-powered one-person business, and the setup is genuinely easy: you can stand up the tools in an afternoon. The setup was never the hard part.

彭博 10 月 25 日发了一篇“如何搭建 AI 驱动的一人公司”的指南，搭建确实不难——一个下午就能把工具跑起来。搭建从来不是难的地方。

Open note阅读短评
The Individual Lab超级个体实验室

Mastercard let an agent pay 9 times on one budget.Mastercard 让一个 agent 用同一个预算付了 9 次款。

A wallet amplifies mis-steered spend; the budget cap, not per-payment review, is your only control.

钱包会放大被带偏的支出；真正的控制不再是逐笔审核，而是你设的预算上限。

Open note阅读短评
The Individual Lab超级个体实验室

Cursor's cheaper bundle meters the model you code with.Cursor 更便宜的捆绑包，对你用来写代码的那个模型按量计费。

Its "all-in-one" plan leaves out the AI most devs actually use.

它的『一体化』套餐，不含多数开发者真正在用的 AI。

Open note阅读短评
The Individual Lab超级个体实验室

n8n's new default survives restarts but drifts off the clockn8n 的新默认重启不丢任务，却也不再对齐整点

Your "every 15 minutes" jobs may fire at new times after you upgrade.

升级之后，你那些「每 15 分钟」的任务可能会在新的时刻触发。

Open note阅读短评
The Individual Lab超级个体实验室

The AI coding leaderboard can't rank its own top models.AI 编码榜单，排不动自己的榜首。

You pay up to 2x for the #1 — and ship the same work.

你为第一名多付最多一倍——交付的活却一样。

Open note阅读短评
Systems Research系统研究

A one-person billion-dollar company? That proves leverage, not solo capability.一个人的十亿美元公司？那证明的是杠杆，不是个人能力。

The pitch sells automation as proof — and skips the one step AI can't take: judgment.

这套说辞把自动化当成证据——却跳过了 AI 接不走的那一步：判断。

Open note阅读短评

All essays and notes全部长文与短评

Base rates first先看基准线

4 figures4 个数字

Before any single story: what the wider data says. Each number keeps its date, its source and what it does not measure.

读任何一个案例之前，先看整体数据怎么说。每个数字都带着日期、来源，以及它没有衡量什么。

  1. 63%

    of C-corps formed through Stripe Atlas in Q2 2026 (to date) had a solo founder — an all-time high.

    2026 年二季度（截至发文）经 Stripe Atlas 设立的 C corp 中，单人创始占 63%，历史最高。

    Does not show:没衡量： Atlas users only; counts incorporations, not surviving businesses.只统计 Atlas 用户；数的是注册公司，不是活下来的生意。

    Stripe — Solo founding is at an all-time highStripe：单人创业比例创历史新高 Platform data平台官方数据

  2. −23% / +19%

    Solo-founded Atlas startups in 2025: median first-six-month revenue fell 23% year over year; the top decile rose 19%. The top-decile-to-median gap grew from about 34× four years earlier to 61×.

    2025 年经 Atlas 设立的单人创始公司：前 6 个月收入中位数同比下降 23%，头部 10% 上升 19%；头部与中位数的差距从四年前约 34 倍拉大到 61 倍。

    Does not show:没衡量： More people start alone; the typical outcome is getting worse, not better.一个人创业的人更多了，但典型结果在变差，不是变好。

    Stripe — Solo founding is at an all-time highStripe：单人创业比例创历史新高 Platform data平台官方数据

  3. 67.5%

    of startups listed on TrustMRR have under $1,000 in payment-verified revenue; 1.0% have more than $1M.

    TrustMRR 收录的项目中，67.5% 经支付端验证的收入不到 1,000 美元；超过 100 万美元的只有 1.0%。

    Does not show:没衡量： Founders opt in, so the sample is self-selected; the marketplace is owned by founder Marc Lou.项目是自愿接入的，样本有自选偏差；平台归独立开发者 Marc Lou 所有。

    TrustMRR — Startup Statistics (daily snapshot)TrustMRR：创业公司统计（每日快照） Payment-verified支付端验证

  4. 426

    OPC (one-person company) communities in China as of May 2026, across 28 provincial regions and 65 cities.

    截至 2026 年 5 月，全国共有 426 个 OPC（一人公司）社区，覆盖 28 个省级行政区、65 个城市。

    Does not show:没衡量： These are communities and incubation spaces, not companies — and not proof of paying customers.这是社区和孵化空间的数量，不是公司数量，更不代表有付费客户。

    Xinhua: 2026 China One-Person Company (OPC) Survey Report released (Fudan University, Yao Jianhua's team)新华网：《2026年中国一人公司（OPC）调研报告》发布（复旦大学姚建华团队） Independently reported媒体报道

Worth learning this week本周值得学的

We read every update from the sources below and kept only what will still matter in a year and teaches you something you can use. The line under each item says what it changes.

下方来源的更新我们都读过，只留下一年后仍有用、读完能多会一件事的条目；每条下面一句写明它改变了你的哪个想法或做法。

3 kept this week, newest first本周留下 3 条，按时间倒序

  1. V2EX · Show & ShareV2EX · 分享创造

    Independently developed a tool for cross-border sellers to test product styles before sampling (free quota available)独立开发了一个出海小工具，帮跨境卖家在打样前测款(有免费额度)

    Testing product visuals before sampling reduces wasted prototyping cost; a copyable pre-sample workflow.

    在打样前用工具测试款式视觉，可减少无效打样成本，是可复用的验品流程思路。

    Hi everyone, I built a tool for cross-border sellers called Museify: https://museify.app — sharing it here on V2EX and would love to hear your feedback. Also noticed that Huiwa seems to be shutting down its service, so former Huiwa users are welcome to give it a try and share your thoughts; I'll keep working to improve my product. Why I built this: It's a painful process for apparel sellers to launch new styles — if you want to try a new color or pattern, you have to get a sample made first (usually 2–5 weeks, costing anywhere from a few dozen to two hundred USD per piece), then shoot photos, list the item, and run ads, only to find out whether it sells. A lot of the time the samples come back and nobody buys them…

    大家好，我做了一个给跨境卖家用的工具 Museify:https://museify.app ,来 V2EX 分享一下，也想听听大家的意见。另外看见绘蛙好像也要停止服务了，原来绘蛙的老用户可以试用一下，给点意见，我努力继续优化我的产品 为什么做这个 做服装的卖家上新款很痛苦：想试一个新颜色、新花型，要先打样(一般 2–5 周，一件几十到两百美元),再拍照、上架、投广告，才知道卖不卖得动。很多款样衣回来才发现没人买…

    Builder community开发者社区Community社区
  2. Failory NewsletterFailory 创业失败案例通讯

    Hack Me Twice两度被黑

    Two breaches in sequence show how trust collapses faster than systems can recover.

    连续两次安全漏洞如何比系统恢复更快地摧毁用户信任。

    Why Zeabur shut down after two consecutive breaches.

    Zeabur为何在连续遭受两次安全漏洞攻击后关闭。

    Cases & post-mortems案例与复盘Media媒体
  3. Acquire.com BlogAcquire.com 博客（公司收购平台）

    The Operations Shift Behind a Turnkey Business Sale一次交钥匙型企业出售背后的运营转型

    Building systems that run without you is what makes a business transferable or scalable.

    让业务脱离创始人独立运转，是可转让、可扩展的关键机制。

    Vincent Serpico spent 16 years building Founders Workshop, a product development company that helped founders turn ideas into working software. By the time he considered selling, the business had built more than 300…

    Vincent Serpico 花了16年时间打造 Founders Workshop——一家帮助创始人将创意转化为可运行软件的产品开发公司。在他考虑出售时，该公司已完成了逾300个……

    Cases & post-mortems案例与复盘Has a stake利益相关

Case board案例看板

10 records10 条记录

Ten records, each with its numbers, the basis for each number, and its sources. Dashed lines mean self-reported. Read the cautionary cases as carefully as the wins.

十条记录，每条写清数字、数字的口径和出处。虚线框表示“自述·未核实”。反面和边界案例要和成功案例一样认真读。

What the labels mean标记是什么意思
  • Payment-verified支付端验证Pulled by a third party from the payment processor through a read-only API (e.g. TrustMRR). Proves revenue, not profit; the verifier can have its own interests.第三方通过支付服务商只读接口拉取（如 TrustMRR）。能证明收入，证明不了利润；验证方本身也可能有利益。
  • Self-reported · unverified自述 · 未核实Only the founder's own posts or interviews.只有当事人自己的帖子或访谈。
  • Independently reported媒体报道Published by a news outlet or research team; the underlying figure may still come from the company.新闻机构或研究团队发布；底层数字仍可能来自公司本身。
  • Platform count平台计数Subscribers, followers or stars shown by a platform: audience, not revenue.平台显示的订阅数、粉丝数或星标数：是受众，不是收入。
  • Official record官方记录Regulator letters, government documents, acquirer announcements.监管信函、政府文件、收购方公告。
  • Third-party estimate第三方估算Modelled by an analyst, tool or industry insiders, not measured.由分析机构、工具或业内人士估算，不是实测。

  1. PostizPostiz

    Team level 团队等级 C
    Software软件

    Nevo David, open-source developer

    Nevo David，开源开发者

    Open-source (AGPL-3.0) social-media scheduler with a paid hosted version

    开源（AGPL-3.0）社交媒体排程工具 + 付费托管版

    Channels渠道 Open source on GitHub (37k+ stars) · Selling to AI agents through its APIGitHub 开源（3.7 万+ 星标） · 通过 API 把 AI 代理当成用户

    Micro team: described himself as a single developer in Nov 2025; has since hired engineers微团队：2025-11 自称单人开发者，之后已招聘工程师

    Facts and their basis事实与口径

    • MRR $282,883 with 8,289 active subscriptions; $1.48M all-time revenue (synced 2026-10-18)MRR 282,883 美元，活跃订阅 8,289 个；累计收入 148 万美元（2026-10-18 同步） [1]source 1来源 1 Payment-verified支付端验证
    • “$14.2k monthly as a single developer” (Nov 2025)“单人开发者月入 1.42 万美元”（2025-11） [2]source 2来源 2 Self-reported · unverified自述 · 未核实
    • Founder says revenue jumped after he started selling to AI agents创始人称开始面向 AI 代理销售后收入跳升 [3]source 3来源 3 Self-reported · unverified自述 · 未核实

    Lesson教训

    AI agents became both users and a distribution channel; open source bought the reach.

    AI 代理同时成了用户和分发渠道；开源换来了触达。

    1. TrustMRR — PostizTrustMRR：Postiz
    2. DEV Community — Nevo DavidDEV 社区：Nevo David
    3. Mixergy — Revenue jumped when he sold to AI agentsMixergy 访谈：开始卖给 AI 代理后收入跳升

    Data as of数据日

  2. Base44Base44

    Team level 团队等级 D
    Software软件Exit退出

    Maor Shlomo, Israel

    Maor Shlomo，以色列

    Bootstrapped vibe-coding app builder for non-programmers

    自筹资金的“氛围编程”应用生成工具，面向不会写代码的人

    Channels渠道 Building in public on LinkedIn and X · Word of mouth在 LinkedIn 和 X 上公开记录开发过程 · 口碑传播

    Solo-owned, not solo-run: eight employees at the sale一人持股，不是一人干活：出售时有 8 名员工

    Facts and their basis事实与口径

    • Sold to Wix for $80M in cash six months after launch (June 2025); $25M of it is a retention bonus for the eight employees上线 6 个月后以 8,000 万美元现金卖给 Wix（2025-06）；其中 2,500 万美元是给 8 名员工的留任奖金 [1]source 1来源 1 Independently reported媒体报道
    • $189,000 profit in May 2025, per the founder's own posts2025 年 5 月利润 18.9 万美元（据创始人自己的帖子） [1]source 1来源 1 Self-reported · unverified自述 · 未核实

    Lesson教训

    The founder sold because scale needed resources he could not grow into organically.

    创始人选择出售，是因为规模化需要的资源靠自然增长跟不上。

    1. TechCrunch — 6-month-old, solo-owned vibe coder Base44 sells to Wix for $80M cashTechCrunch：成立 6 个月、一人持股的氛围编程工具 Base44 以 8,000 万美元现金卖给 Wix

    Data as of数据日

  3. Pieter Levels (Photo AI, Nomads and others)Pieter Levels（Photo AI、Nomads 等）

    Team level 团队等级 A
    Software软件

    Pieter Levels, Netherlands

    Pieter Levels，荷兰

    A portfolio of small paid web apps

    一组小型付费网页应用

    Channels渠道 Personal audience on X · Public revenue posts on his own blogX 上的个人受众 · 在自己博客公开收入

    Solo纯一人

    Facts and their basis事实与口径

    • Photo AI: $105,000/month revenue, $80,000/month profit (Mar 2026)Photo AI：月收入 10.5 万美元，月利润 8 万美元（2026-03） [1]source 1来源 1 Self-reported · unverified自述 · 未核实
    • “$10M/y in revenue and investment gains” — the headline figure includes investment gains (Sep 2026)“年 1,000 万美元收入加投资收益”——这个总数包含投资收益（2026-09） [2]source 2来源 2 Self-reported · unverified自述 · 未核实

    Lesson教训

    The moat is attention, low costs and capital, not the code. Separate operating revenue from investment gains.

    护城河是注意力、低成本和资本，不是代码。要把经营收入和投资收益分开看。

    1. levels.io — Photoai.com is a 40,870 line index.php…levels.io：Photoai.com 是一个 40,870 行的 index.php……
    2. levels.io — I passed $10M/y in revenue and investment gains…levels.io：我的收入加投资收益已超过每年 1,000 万美元……

    Data as of数据日

  4. Lenny's NewsletterLenny's Newsletter

    Team level 团队等级 B
    Media & newsletters媒体与 Newsletter

    Lenny Rachitsky, United States

    Lenny Rachitsky，美国

    Paid newsletter, podcast and sponsorships

    付费 Newsletter + 播客 + 赞助

    Channels渠道 Substack · PodcastSubstack · 播客

    No full-time employees; about 10 contractors0 名全职员工，约 10 名合同工

    Facts and their basis事实与口径

    • Over 1,200,000 subscribers shown on the Substack page (Oct 2026)Substack 页面显示 120 万+ 订阅（2026-10） [1]source 1来源 1 Platform count平台计数
    • No full-time staff, about 10 contractors worldwide (his account, Apr 2026)无全职员工，全球约 10 名合同工（本人讲述，2026-04） [2]source 2来源 2 Self-reported · unverified自述 · 未核实

    Lesson教训

    Hand off execution, keep the judgment. Subscriber counts measure audience, not income.

    执行交给外包，判断留给自己。订阅数衡量的是受众，不是收入。

    1. Lenny's Newsletter (Substack)Lenny 通讯（Substack）
    2. First Round Review — Reluctantly InfluentialFirst Round Review：不情愿的影响力

    Data as of数据日

  5. Dan KoeDan Koe

    Team level 团队等级 ?
    Media & newsletters媒体与 Newsletter

    Dan Koe, writer and course creator, United States

    Dan Koe，写作者与课程作者，美国

    Newsletter, courses and premium guides, plus a software product

    Newsletter、课程和付费指南，另有软件产品

    Channels渠道 Long-form posts on X · NewsletterX 长文 · Newsletter

    Content team size not disclosed; co-founder of the software product Eden内容业务团队规模未公开；软件产品 Eden 为联合创办

    Facts and their basis事实与口径

    • Awarded $250,000 in X's long-form article contest (Feb 2026)在 X 长文大赛中获得 25 万美元奖金（2026-02） [1]source 1来源 1 Independently reported媒体报道

    Lesson教训

    A platform prize is paid once. Reach is the ticket; the funnel is the business.

    平台奖金只发一次。流量是门票，漏斗才是生意。

    1. NBC News — X awards $1 million prize…NBC 新闻：X 颁出 100 万美元奖金……
    2. thedankoe.comthedankoe.com（Dan Koe 个人网站）

    Data as of数据日

  6. PolsiaPolsia

    Team level 团队等级 B
    Software软件Cautionary反面与边界

    Ben Cera (Victor-Benjamin Broca), United States

    Ben Cera（Victor-Benjamin Broca），美国

    AI-agent platform that runs small companies for its users

    让 AI 代理替用户经营小公司的平台

    Channels渠道 Founder's posts on LinkedIn and X; press coverage创始人在 LinkedIn、X 发帖；媒体报道

    No employees; legal and infrastructure handled by partner firms (his “virtual team”)零员工；法务和基础设施交给合作公司（他称为“虚拟团队”）

    Facts and their basis事实与口径

    • Announced a $30M raise at a $250M valuation (May 2026)宣布以 2.5 亿美元估值融资 3,000 万美元（2026-05） [1]source 1来源 1 Independently reported媒体报道
    • Revenue run rate is the founder's own figure; Fortune's reporter said he had not verified it (Mar 2026)营收年化数字来自创始人本人；《财富》记者明言未核实（2026-03） [2]source 2来源 2 Self-reported · unverified自述 · 未核实
    • First-month churn “around 50%”, in the founder's words (May 2026)首月流失率“约 50%”，创始人原话（2026-05） [3]source 3来源 3 Self-reported · unverified自述 · 未核实

    Lesson教训

    AI can set up a company; it cannot set up demand.

    AI 能把公司搭起来，搭不出需求。

    1. Pulse 2.0 — Polsia: $30 Million At $250 Million ValuationPulse 2.0：Polsia 以 2.5 亿美元估值融资 3,000 万美元
    2. Fortune — The one-person unicorn: Myth, miracle, or the future of startups?《财富》：一人独角兽——神话、奇迹，还是创业的未来？
    3. Mixergy — interview with Ben CeraMixergy：Ben Cera 访谈

    Data as of数据日

  7. MedviMedvi

    Team level 团队等级 C
    Commerce电商Cautionary反面与边界

    Matthew Gallagher, Los Angeles

    Matthew Gallagher，洛杉矶

    Direct-to-consumer GLP-1 telehealth marketing

    直接面向消费者的 GLP-1 减重药远程医疗营销

    Channels渠道 Website and online marketing (the FDA letter concerns website claims) · Email marketing (named in a pending California anti-spam class action)官网与线上营销（FDA 警告信针对的就是官网宣传内容） · 邮件营销（被一起待决的加州反垃圾邮件集体诉讼点名）

    Two full-time people (the founder and his brother); started with no employees and more than a dozen AI tools, per reports2 名全职（创始人和弟弟）；据报道起步时没有员工，靠十几种 AI 工具

    Facts and their basis事实与口径

    • $401M in sales in its first full year, per The New York Times (Apr 2026); not an audit首个完整年度销售额 4.01 亿美元，据《纽约时报》报道（2026-04），非审计数据 [1]source 1来源 1 Independently reported媒体报道
    • FDA warning letter to MEDVi, LLC dated 2026-02-20 over its website claimsFDA 于 2026-02-20 就网站宣传内容向 MEDVi, LLC 发出警告信 [2]source 2来源 2 Official record官方记录
    • The Times later added an editor's note on the FDA letter and a pending class action《纽约时报》事后补充编者按，写明 FDA 警告信和一起待决集体诉讼 [3]source 3来源 3 Independently reported媒体报道

    Lesson教训

    Outsourcing moves the work, not the responsibility. A headline is not due diligence.

    外包能转移工作量，转移不了责任。一篇报道不等于尽职调查。

    1. PYMNTS — The one-person billion-dollar company is here (citing NYT)PYMNTS：十亿美元的一人公司来了（援引《纽约时报》）
    2. FDA Warning Letter — MEDVi, LLC dba MEDVi (MARCS-CMS 721455)美国 FDA 警告信：MEDVi, LLC（MARCS-CMS 721455）
    3. Futurism — New York Times edits Medvi articleFuturism：《纽约时报》修改 Medvi 报道

    Data as of数据日

  8. AI-only YouTube channels纯 AI 内容 YouTube 频道

    Team level 团队等级 ?
    AI contentAI 内容Cautionary反面与边界

    278 channels among the world's top YouTube channels

    全球热门 YouTube 频道中的 278 个

    Mass-produced AI video monetised through ad revenue share

    批量生产 AI 视频，靠广告分成变现

    Channels渠道 YouTube recommendationsYouTube 推荐流

    Group case; operators mostly anonymous群像；运营者大多匿名

    Facts and their basis事实与口径

    • About $117M a year in combined revenue across 278 AI-only channels (estimate cited by The Guardian, Dec 2025)278 个纯 AI 频道年收入合计约 1.17 亿美元（《卫报》2025-12 引用的估算） [1]source 1来源 1 Third-party estimate第三方估算
    • YouTube renamed its “repetitious content” policy “inauthentic content”, clarifying that mass-produced, repetitive videos cannot be monetised (15 Jul 2025)YouTube 把“重复内容”政策更名为“不真实内容”，明确批量制作、重复的视频不能变现（2025-07-15） [2]source 2来源 2 Official record官方记录

    Lesson教训

    What gets demonetised is volume, sameness and deception — not AI as such. Rented reach can be switched off.

    被取消变现的是量产、雷同和误导，不是 AI 本身。租来的流量随时可以被关掉。

    1. The Guardian — More than 20% of videos shown to new YouTube users are ‘AI slop’《卫报》：向 YouTube 新用户展示的视频中超过 20% 是“AI 垃圾内容”
    2. YouTube Help — YouTube channel monetization policiesYouTube 帮助中心：频道获利政策

    Data as of数据日

  9. Liu Xiaopai (Raphael AI and other tools)刘小排（Raphael AI 等工具）

    Team level 团队等级 D
    Software软件Cautionary反面与边界

    Liu Xiaopai, China; former product director

    刘小排，中国；前大厂产品总监

    AI products for overseas users, paid tiers plus some ads; also co-runs courses with a paid community

    面向海外用户的 AI 产品，靠用户付费和少量广告；另与付费社群合作出课程

    Channels渠道 Customer acquisition methods not disclosed in detail · Personal audience on X and in Chinese communities获客方式未详细披露 · X 和中文社区里的个人影响力

    Often called a one-person company; he says the company has six or seven people常被称为“一人公司”；本人说公司“只有六七个人”

    Facts and their basis事实与口径

    • Close to ¥10M revenue in 2024 from user payments and some ads, as he told an interview show (InfoQ via 36Kr, Aug 2025)本人在访谈中称 2024 年收入接近 1,000 万元，来自用户付费和少量广告（InfoQ，36氪转载，2025-08） [1]source 1来源 1 Self-reported · unverified自述 · 未核实
    • “Revenue could support thirty to fifty people at market rates, but the company has only six or seven” (Sep 2025)“按市场价格可以养三五十个人……公司只有六七个人”（2025-09） [2]source 2来源 2 Self-reported · unverified自述 · 未核实

    Lesson教训

    What matters is a short decision chain, not a headcount of one.

    关键是决策链短，不是人数只有一个。

    1. 36Kr (republished from InfoQ)36氪（转载 InfoQ）
    2. Tencent News · Crossing interview腾讯新闻 · 十字路口 Crossing 访谈

    Data as of数据日

  10. Solo AI comic-drama creators (China)AI 漫剧个人创作者（中国）

    Team level 团队等级 ?
    AI contentAI 内容Cautionary反面与边界

    Creators on Chinese short-video platforms

    国内短视频平台上的创作者

    AI-made serial comic dramas paid by platform revenue share

    用 AI 制作连载漫剧，靠平台分成

    Channels渠道 Short-video and mini-drama platforms短视频和微短剧平台

    Group case; many work alone with AI tools群像；很多人一个人加 AI 工具完成全流程

    Facts and their basis事实与口径

    • Industry estimate: fewer than 1% of top creators earn ¥100,000 to nearly ¥1M a month; most creators earn under ¥10 a month (Beijing Business Today, Mar 2026)据业内估算，不足 1% 的头部创作者月入十万元到近百万元，绝大多数创作者单月收益不足 10 元（北京商报，2026-03） [1]source 1来源 1 Third-party estimate第三方估算

    Lesson教训

    When the barrier drops to the floor, the distribution gets steeper, not flatter.

    门槛降到地板，收入分布只会更陡，不会更平。

    1. Beijing Business Today (via JWView): one-person companies behind the comic-drama boom北京商报（中新经纬转载）：漫剧爆火下的“一人公司”

    Data as of数据日

Reviewed every two weeks; numbers are snapshots on the date shown.

每两周人工复核一次；数字是所示日期的快照。

Rule-change log规则变更记录

8 changes8 条变化

Platform, tax and content rules that change what a one-person business keeps. Official sources only; newest first.

改变一人公司能留下多少的平台、税务与内容规则。只用官方来源，按时间倒序。

  1. Global全球Platform平台

    Apple苹果 · App Review Guidelines updatedApp 审核指南更新

    New guideline 2.3.14 on creative assets; new 2.5.19 on hidden, dormant or undocumented features; 3.1.1(a) clarifies where external purchase links may point.

    新增 2.3.14（创意素材）和 2.5.19（隐藏、休眠或未说明的功能）；3.1.1(a) 澄清外部购买链接可以指向哪里。

    Affects:影响谁： Independent iOS developers独立 iOS 开发者

    Apple Developer News苹果开发者新闻

  2. UK英国Tax税务

    HMRC英国税务海关总署 · Making Tax Digital for Income Tax widens所得税数字化申报（MTD）扩围

    Mandatory since 6 April 2026 for sole traders and landlords with income over £50,000; extends to over £30,000 from 6 April 2027 and over £20,000 from April 2028. Digital records and quarterly updates through compatible software.

    自 2026-04-06 起，收入超过 5 万英镑的个体经营者和房东必须使用；2027-04-06 起扩大到超过 3 万英镑，2028 年 4 月起到超过 2 万英镑。需保存数字记录，并用合规软件按季度报送。

    Affects:影响谁： UK sole traders英国个体经营者

    GOV.UK — Sole traders and landlords earning more than £30,000 urged to act now英国政府：年收入超过 3 万英镑的个体经营者和房东请立即行动

  3. China中国Policy政策

    MIIT (China)工业和信息化部 · AI SME startup support plan (2026–2028) issued《人工智能中小企业创业支持计划（2026-2028年）》印发

    A three-year national support plan for AI small-business startups (MIIT General Office document Qiye [2026] No. 31, dated 29 August 2026).

    面向人工智能领域中小企业创业的三年支持计划（工信厅企业〔2026〕31号，2026-08-29 成文）。

    Affects:影响谁： AI founders and small companies in China国内 AI 创业者和中小企业

    Ministry of Industry and Information Technology of China工业和信息化部

  4. China中国Content内容

    Cyberspace Administration of China中央网信办 · Second phase of the campaign against AI misuse“清朗·整治AI应用乱象”专项行动第二阶段

    Targets AI-made false information, violent or vulgar content, impersonation and harm to minors; platforms told to clear illegal AI content, and violating accounts and apps to be dealt with.

    聚焦用 AI 制作发布虚假信息、暴力低俗内容、假冒仿冒他人、侵害未成年人权益；督促平台集中清理 AI 生成的违法不良信息，依法查处违规账号和应用。

    Affects:影响谁： AI content accounts on Chinese platforms国内平台上的 AI 内容账号

    Xinhua (republished from the Cyberspace Administration of China)新华网（转载中国网信网）

  5. EU欧盟Tax税务

    European Union欧盟 · €150 duty exemption for small parcels abolished取消 150 欧元以下小包裹免关税

    From 1 July 2026 a temporary €3 customs duty applies per item on low-value consignments (up to €150) from outside the EU, until 1 July 2028; a €2 Union handling fee applies from 1 November 2026.

    自 2026-07-01 起，从欧盟以外进口的低价值包裹（150 欧元以下）按件征收 3 欧元临时关税，至 2028-07-01；2026-11-01 起另收 2 欧元的欧盟处理费。

    Affects:影响谁： Cross-border e-commerce sellers shipping into the EU向欧盟发货的跨境电商卖家

    EU Taxation and Customs Union欧盟税务与海关联盟

  6. China中国Tax税务

    State Council (China)国务院 · Platforms start reporting sellers' and workers' income to tax authorities互联网平台开始报送平台内经营者和从业人员涉税信息

    Under State Council Decree No. 810 (published and effective 20 June 2025), platforms report the identity and income information of sellers and individual service providers; first reports from 1 October 2025, then within the month after each quarter.

    依据国务院令第 810 号《互联网平台企业涉税信息报送规定》（2025-06-20 公布并施行），平台报送平台内经营者和从业人员的身份与收入信息；2025-10-01 起首次报送，此后每季度结束次月内报送。

    Affects:影响谁： Sellers and creators earning through Chinese platforms通过国内平台获得收入的商家和创作者

    Chinese Government Portal · State Council Decree No. 810中国政府网 · 国务院令第810号

  7. China中国Content内容

    Cyberspace Administration of China and three other agencies国家网信办等四部门 · AI-generated content labelling rules take effect《人工智能生成合成内容标识办法》施行

    AI-generated or synthesised content must carry explicit or implicit labels; users who publish such content must declare it and use the platform's labelling function (Article 10).

    AI 生成合成内容须添加显式或隐式标识；用户发布此类内容，应主动声明并使用平台提供的标识功能（第十条）。

    Affects:影响谁： Anyone publishing AI content in China在国内发布 AI 内容的所有人

    Cyberspace Administration of China中国网信网

  8. Global全球Platform平台

    YouTubeYouTube · “Repetitious content” becomes “inauthentic content”“重复内容”政策更名为“不真实内容”

    YouTube clarified that repetitive or mass-produced content is ineligible for monetisation in the Partner Programme.

    YouTube 澄清：重复或批量制作的内容不能在合作伙伴计划中变现。

    Affects:影响谁： Creators monetising on YouTube在 YouTube 上变现的创作者

    YouTube Help — channel monetization policiesYouTube 帮助中心：频道获利政策

Questions we keep asking我们持续追问的问题

  1. Who verified the revenue: a payment processor, a regulator, a reporter — or only the founder?这个收入数字是谁验证的：支付服务商、监管机构、记者，还是只有创始人自己？
  2. Is the distribution owned (an email list, a community, a product people search for) or rented from an algorithm?分发是自己的（邮件名单、社群、有人主动搜索的产品），还是租来的算法流量？
  3. If the platform changes its cut, its payout rules or its content policy tomorrow, how much of the margin survives?如果平台明天改抽成、改结算规则或改内容政策，利润还剩多少？
  4. Which work went to AI, which to contractors, and which can only the founder do?哪些活交给了 AI，哪些交给了外包，哪些只能创始人自己做？
  5. Is the median solo business getting better, or only the top decile?变好的是中位数的一人公司，还是只有头部那 10%？
  6. After the subsidies and the community launch events, are there paying customers?补贴和社区挂牌之后，有没有付钱的客户？
  7. What would change our view: an audited one-person company at $1B in revenue or valuation, or base rates that keep widening?什么证据会让我们改变判断：一家经审计、营收或估值达到 10 亿美元的一人公司，还是持续拉大的分化数据？

Concepts behind this page相关概念

  • From possibility to trade-off从可能到取舍

    Moving from seeing a possibility to deciding on a trade-off: be clear about your goal, your conditions and the cost you are willing to bear.

    从看见可能到决定取舍：明确目标、条件和愿意承担的代价。

  • Results alone don't prove capability能力不能由结果单独证明

    Wealth, follower counts and degrees of automation may be results of practice, but on their own they do not prove judgment, creativity or autonomy.

    财富、粉丝数和自动化程度可能是实践结果，但不能单独证明判断力、创造力或自主性。

What we read, and why我们看哪些来源，为什么

Public sources only. Headlines and the sources' own short summaries are translated before publishing; every item links to the original.

只用公开来源。标题和来源自带的简短摘要在发布前翻译好；每一条都链接原文。

Show all 20 sources and their status展开全部 20 个来源与状态

Cases & post-mortems案例与复盘

Platforms & payments平台与支付

Policy, tax & enforcement政策、税务与执法

Industry & media行业与媒体

Builder community开发者社区

How this page works这个页面怎么来的

We gather public sources several times a day and keep only what passes two gates: still useful in a year, and you can think or do one new thing after reading it. Kept items are translated before publishing, with one line on what each changes, and the page is rebuilt as plain HTML, so it reads entirely in English or entirely in Chinese. Each item links to the original; WAKENEX readings are our own essays. Machine-readable copies are listed on the page for AI agents.

我们每天数次汇集公开来源，只留下过了两道门的条目：一年后仍有用，读完能多想清楚或多做对一件事。留下的条目在发布前翻译好，每条附一句它改变了你的哪个想法或做法，再重新生成这个静态页面，所以整页要么全是中文，要么全是英文。每条都链接原文；“WAKENEX 解读”是我们自己的长文与短评。给机器读的版本列在“给 AI 代理”页。

For AI agents给 AI 代理 · All topics全部主题