非常有意思。我本来正打算写这个。 你可以说,过去所有的企业软件基本都只是“还行”,因为客户不得不用它来运营业务,所以真正创造价值的是客户自己的员工,而不是软件本身。因此,应用软件市场一年大概也就 2000 亿美元规模。 而对 AI SaaS 的乐观观点在于: 它终于真正兑现了 SaaS 本应实现的价值。 Super interesting. Was going to write about this. You could argue all enterprise sw was done "ok" because the customer needed to operate it. So value was dependent on their staff. Hence apps is maybe 200b per yr. A bull case for ai SaaS is that it finally delivers on SaaS.
目录
观点原文
非常有意思。我本来正打算写这个。
你可以说,过去所有的企业软件基本都只是“还行”,因为客户不得不用它来运营业务,所以真正创造价值的是客户自己的员工,而不是软件本身。因此,应用软件市场一年大概也就 2000 亿美元规模。
而对 AI SaaS 的乐观观点在于:
它终于真正兑现了 SaaS 本应实现的价值。
Super interesting. Was going to write about this. You could argue all enterprise sw was done "ok" because the customer needed to operate it. So value was dependent on their staff. Hence apps is maybe 200b per yr. A bull case for ai SaaS is that it finally delivers on SaaS.
ABAB AI 解读
一句话核心判断
传统企业软件之所以长期停留在“2000 亿美元级别”,不是市场不大,而是软件本身并没有真正创造价值;AI SaaS 的潜力在于,它第一次让软件本身成为“产出者”,而不是工具。
一、为什么说过去的企业软件只是“done ok”
这句话不是贬低 SAP、Oracle、Salesforce,而是在指出一个事实:
传统企业软件的本质是:
- 流程工具
- 记录系统
- 管理界面
真正创造价值的仍然是:
- 员工操作
- 人工决策
- 人工执行
所以软件的价值函数是:软件 × 人的能力
而不是:软件本身直接产出结果
二、“客户需要它才能运转”是一个反讽
“the customer needed to operate it” 表面是刚需,实则是限制:
- 你必须雇人来用它
- 你必须培训人
- 你必须为低效流程买单
这意味着:软件的上限,被人的规模与效率锁死了。
三、为什么应用软件市场只有 ~200B/年
作者给出的隐含逻辑是:
如果软件只是:帮人更好地干活
那么市场规模 ≈ 企业愿意为“效率提升工具”付的钱
这自然是一个:
- 成熟
- 增长有限
- 高度竞争的市场。
四、AI SaaS 的“牛市逻辑”到底是什么
关键一句是:
“AI SaaS finally delivers on SaaS.”
意思是,SaaS 最初承诺的是:
- 软件即服务
- 软件替你做事
但过去几十年:软件只是“让你更好地做事”
而 AI SaaS 的本质变化是:软件开始真正“替代一部分人力产出”。
五、从“工具”到“劳动力”的跃迁
AI SaaS 改变的是价值锚点:
传统 SaaS 定价锚点:
- seat 数
- 模块数
AI SaaS 的潜在锚点:
- 任务完成量
- 结果交付
- 产出规模
这意味着:软件第一次可以按“创造的价值”而不是“使用的权限”收费。
六、为什么这可能是 100x 市场的起点
一旦软件:
- 能独立完成任务
- 能持续产出
- 能减少用人成本
那么企业愿意支付的,不再是:IT 预算的一小块
而是:原本属于人力成本、外包成本、服务成本的那一大块。这就是市场从 200B → 数万亿 的理论来源。
总结一句话
这段话真正想表达的是:过去的软件只是“看起来在工作”,而 AI SaaS 是第一次真的在“工作”。
如果成立,那么:
- 过去的 SaaS 估值模型会失效
- 市场天花板会被重新打开
- “软件公司 vs 服务公司”的界线将开始消失
相关观点
- 关于招聘,我学到了一件事:(几乎)每个人都在找机会。 (几乎)每一个人都希望往上走、承担更多责任、赚更多钱等等,只是需要你多花一些功夫才能触达到他们。 推论是:我团队里的每一个人其实也都在“看机会”,所以我必须持续投入精力,努力把他们留下来。 One thing I learned about recruiting: (almost) everyone is looking. (Almost) every single person is looking to move up, take on more, make more $, etc. They just take more work to get to. Corollary: Everyone on my teams were also looking, so I needed to work to keep them!Nick Mehta · 2025/12/23
- 对于一家 B2B 创业公司来说,最重要的决策之一就是选择正确的早期客户。因为当你开始为这些客户构建产品时,你的公司发展路径会被他们深深影响。“谁愿意买就卖给谁”其实是非常危险的。 One of the biggest decisions for a B2B startup is picking the right early clients. There is path dependency from those customers as you build your company for them. Selling to anyone who wants to buy can be dangerous.Nick Mehta · 2025/12/11
- 创业圈里最糟糕的建议之一,就是告诉别人“最好的产品总是会赢”。有一整座创业公司的“墓地”都死在这句建议上。 “好产品”是必要条件,但远远不够。 One of the worst pieces of advice in startups is telling people that the best product always wins. There is a graveyard of companies that followed that advice. "Great products" are necessary but not sufficient.Nick Mehta · 2025/12/09
- 投资人 @gokulr 最近发了一篇很好的帖子,谈到创业者/CEO 每周给公司写邮件的重要性。我自 2006 年以来,一直每周日在晚上给我所管理的团队/公司发送一封邮件。以下是我从中得到的一些经验: 持续性: 无论发生什么,每周都要写。领导者常常低估“可预测性”对团队的重要性。我记得有一次我生病了,邮件晚了一天才发,我们有位员工给我发短信问:“一切都还好吗?” 简单的格式: 多年来,我不断调整内容结构,最终定下了这样的格式: * 本周你需要知道的 3 件事 * 本周推荐的一篇文章 * 本周的公司价值观 * 我下周的计划 * 我工作之外正在关注的东西 下面分别展开来说: 必须知道的事项: 在任何公司——无论大小——每周都会发生很多事。我会强迫自己思考:有哪些 3 件最重要的事情,是我希望每一位员工都知道的? * 有时候是关于新版本的发布; * 有时候是即将到来的活动; * 有时候是董事会会议里发生的内容。 本周推荐文章: 我非常喜欢“一起学习”这个理念。我们在 Slack 上会分享很多内容,但我希望强调一篇所有人都应该读的博客或社交媒体帖子。 本周价值观: 我们有 5 条公司价值观,而且我们非常重视它们。我会在每次邮件中挑一条出来,讨论我们如何在日常工作中真正践行这一价值观。 本周计划: 我的风格是高度透明。我的日程表一直对团队公开。但我希望大家还能看到一个我本周重点关注事项的总结——尤其是我将要会见的客户。 工作之外: 这里可以分享我和孩子做的事情、我正在读的书、我去看的演唱会等等。向团队展现你的“人性面”有着极大的价值。 自我反思: 这个过程对我来说最有价值的一点,是它迫使我回顾自己上一周做了什么、接下来要做什么。有时候我开始写邮件时会发现:“哇,我好像什么都没完成——根本没什么可写的。”那一瞬间就能把我敲醒。 个性化: 我真的把它做成了属于自己的东西。标题用的是我自创的文字梗 ——“Mehtaphysical Musings(Mehta式形而上沉思)”。里面充满各种宅味的物理学和哲学类比,我敢肯定有些人看到都会 cringe(笑)。但那就是 100% 的我。我认为这非常重要。 Investor @gokulr recently had a great post about the value of weekly founder/CEO emails to the company. I've sent a Sunday night email to the various teams/companies I've run every week since 2006. Here are some things I learned: 1. Consistency: Do it every week, no matter what. It's hard for leaders to appreciate how important predictability is for teams. I remember one time where my email came out a day later because I was sick and one of our employees texted and said "is everything ok?" 2. Simple Format: Over the years, I iterated on the content. Eventually I settled on: * 3 Things You Need to Know * 1 Post of the Week * 1 Company Value of the Week * My Plans for the Week * What I'm Into Outside of Work Clicking into each of those: 3. Need to Know: There is so much that happens in any company - big or small. I forced myself to think about what are the 3 most important things that I want every employee to know. Sometimes it was about a new release. Or an upcoming event. Or what happened in the Board Mtg. 4. Post of the Week: I love the idea of learning together. We'd share a ton on Slack, but I wanted to emphasize one blog or social post that everyone should read. 5. Value of the Week: We had 5 company values and were passionate about them. I'd try to talk about one each time and how we can live them. 6. Plans for the Week: My style is extreme transparency. My calendar was always shared. But I wanted folks to get a summary of what I was focused on - especially customers I was meeting. 7. Outside of Work: This could be things I'm doing with my kids, a book I'm reading, a concert I went to, etc. There is so much value in showing your team your human side. 8. Reflection: The best part of this process for me was that I was forced to reflect on what I did in the last week and what I was going to do. Sometimes, I'd start my email and be like "wow I am not getting a lot done - there's nothing to talk about." That would wake me up. 9. Personalization: I really made it my own. The title was my play on words - "Mehtaphysical Musings." It would be full of nerdy physics and philosophy analogies that I'm sure would cause people to cringe. But it was 100% me. I think that goes a long way.Nick Mehta · 2025/12/01