⛏️创客淘金
AI金融数据DaaSB2C订阅+B2B数据授权落地可行性

打造首个AI原生金融数据DaaS,年营收达7位数

Growing the first AI-native financial DaaS to a 7-figure ARR

Fiscal.ai · AI金融数据DaaS · 未公开 · 中7位数ARR

收入规模
$7,000,000 / 年
团队规模
未公开
启动速度
未公开
复刻难度
★★★★☆
访谈亮点
  • 核工程跨界背景发现金融数据中端空白市场
  • 🎯零付费广告靠产品自然增长积累数十万C端用户
  • 🚀C端订阅+B端数据授权双模式支撑中7位数ARR
  • 💡用丰田Kaizen持续迭代框架打磨高门槛数据产品

瞄准金融数据产品中端空白市场,靠AI原生架构解决异构数据清洗痛点,零付费广告做到中7位数ARR,核心壁垒是海量金融数据归一化处理能力。

Fiscal.ai

从核工程到金融科技

From nuclear to fintech

我是一名工程师,之前从事核电和水电相关工作,我的目标一直是发明对地球有益的下一代科技产品,现在有机会的话我也依然在朝着这个方向努力。后来我开始痴迷于市场、投资,研究古今优秀企业家的所有内容,我听完看完了他们产出的所有资料:每一封股东信、每一次访谈、每一本书。我发现普通人不需要巨额资本,就能在公开市场轻松持有这些优秀企业的少量股权,这简直像魔法一样。但在我做研究的过程中市场空白,我发现相关产品存在巨大的中间层空白:一端是昂贵且门槛极高的金融数据终端,另一端是免费、难用、满是广告、完全不适合严肃研究的平台,中间没有任何合适的产品。所有上市公司的财报都是公开免费的,为什么要给所有人打造一个干净专业的聚合体验就这么难?所以我从零开始打造了首个AI原生金融数据公司,解决传统大规模人工聚合数据成本高、延迟高的痛点。现在已有数十万用户信任Fiscal.ai,超过50家企业在自己的产品中使用我们的数据,我们当前年营收达中7位数,预计6个月内就能突破8位数。

I am an engineer who worked professionally in nuclear and hydropower. My aim was to invent the next big thing in technology that was good for the planet. It still is — when I get the chance. But I became obsessed with markets, investing, and studying some of the great entrepreneurs of the past and present. I listened to and read everything they produced. Every shareholder letter, every interview, every book. It felt like magic that, learning from these entrepreneurs and operators in real time, I could easily own a small piece of equity in their businesses easily on the public market without a huge sum of capital. And along the way, I found a huge missing middle in the products for my research. Expensive and exclusive financial data terminals on one end of the spectrum. At the other, free, clunky, ad-filled, and unserious platforms. Nothing in the middle. The financial statements of every public company are freely available! Why was aggregating and delivering a clean, professional experience for everyone so prohibitively difficult? So I started building the first AI-native financial data company from the ground up, to fix problems with the massive dataset aggregated manually with huge costs and latency. Hundreds of thousands now trust Fiscal.ai, and over 50 enterprises now leverage our data in their products. We're making mid-seven figures and we expect to hit eight figures within six months.

?

Who is it for?

  • 有工程背景的资深开发者
  • 熟悉金融数据领域的创业者
  • 能接受长期打磨数据产品的团队

Not for

  • 追求快速变现的短平快创业者
  • 无数据处理能力的纯营销团队
  • 缺乏合规资源的小团队

永远持续改进

Continuous improvement forever

我们最初的MVP重度依赖第三方接入所有必要数据,前端使用React和Next开发,后端集成了多款大语言模型,基于Python和Cloudflare搭建。我技术栈发现做数据(DaaS)产品的周期比普通软件(SaaS)产品长得多,而我们同时在做两类产品。但我们没有等产品完美才上线迭代,产品永远不可能完美,你要做的是永远追求持续改进。我们一直在调整方向解决问题,在做实验、领先市场和聚焦核心问题、倾听用户反馈之间寻找健康的平衡。我从日本丰田的「改善(Kaizen)」框架中获得了很多灵感:哪怕每次只做微小的改进,也要一直坚持优化。

The initial product relied heavily on third parties to pipe in everything needed for a useful MVP. We use React and Next for the frontend. The backend uses many LLMs, Python, and Cloudflare. I found out that building data (DaaS) products takes a lot longer than software (SaaS) products. And we're doing both at the same time. But we didn't wait long to launch and iterate. We didn't wait until it was perfect. It's never perfect. You chase continuous improvement forever. We're constantly pivoting to solve problems. Trying to find a healthy balance between experiments and staying ahead of the market, but also staying grounded and focused on our core problems and customer feedback. I am inspired by Japanese manufacturing companies like Toyota and their "Kaizen" framework, which means continuously improving, even in small chunks, all the time!

⚙️

技术与运营要点

MVP阶段优先用第三方数据源快速验证需求,后续逐步自研数据管道,Kaizen小步迭代模式非常适合高门槛数据类产品,避免一次性投入过大。

产品驱动增长+直销模式

Product-led growth and direct sales

所有你认为对产品有效的方法都可以试,然后把资源加倍投入到真正有效的方向上。我们一直坚持产品驱动增长,这对自助式B2C业务效果极好,至今我们没有在广告上花过一分钱。但我们同时也做B2B业务,所以我们也在逐步拓展直销渠道。早期有人告诉我不可能同时做两类业零广告投入务,我部分同意这个观点,但如果两类用户用的是同一款产品,只是销售模式不同,就完全可行,不存在不可逾越的障碍。

Try everything you believe will work for your product and triple down on what works. We've always leaned into product-led growth, which works great for self-serve B2C. We haven't wasted any money on ads yet. But we also do B2B, so we're increasingly doing direct sales. Early on, people told us we couldn't do both. I partially agree, but you can if the product is the same for both categories. It is just a different sales motion, which is not a dealbreaker.

📈

增长策略组合

同一款产品适配C端自助订阅和B端数据授权两种销售模式,不需要为两类用户单独开发产品,大幅降低了研发冗余。

先行动起来

Just start

我的建议很简单:先动手做,先找到几个需要你帮他们解决问题的客户,其他事情都不重要。相信你的直觉,早期阶段当CEO就是要在信息不全的情况下做决策,依靠你的直觉。同时不要因为决策原则短期成绩过度兴奋,也不要因为挫折过度沮丧,说起来容易做起来难。

My advice is simple: Just start. And get a few customers who want you to help them solve their problems. Nothing else matters yet. Trust your intuition. Being the CEO at the early stages is all about making decisions with incomplete information and relying on your gut. And don't get too high or too low. Easier said than done.

🧠

核心创业心态

早期不要等所有信息完备再启动,先找到3-5个愿意付费解决问题的种子客户,验证需求后再投入全部资源,避免自嗨式开发。

未来规划

What's next?

产品层面我们有宏大的目标,我们很快就要达到8位数ARR,之后我下阶段目标们会想办法再翻一倍。个人层面我还没有明确的规划,说起来有点傻,但这是实话。你可以在推特上关注我,也可以访问Fiscal.ai了解我们的产品。

Strategically and for our products, we have big goals as a company. We're approaching eight figures in ARR and then we'll figure out how to double again. Personally, I am not sure, which sounds a bit silly, but it's the truth. You can follow along on Twitter. And check out Fiscal.ai!

📅

项目里程碑

启动阶段 · 发现金融数据中端空白,打造AI原生金融数据平台

MVP阶段 · 基于第三方数据源搭建最小可用产品,快速上线迭代

增长阶段 · 积累数十万C端用户,拓展50家B端数据授权客户

当前阶段 · 达成中7位数ARR,6个月内冲击8位数ARR

用户评论与创作者回复

已过滤 spam、低价值附和与重复内容 · 原始 106 条 · 展示 7 条 · 创作者回复 0

7
网友RovaAI
👍 1
点透壁垒

The key technical insight is subtle but important: the aggregation, normalization, and AI-native querying is what creates the value. Financial data is notoriously messy, the engineering work of cleaning and normalizing this at scale is the actual barrier to entry.

核心技术洞察非常精准:数据聚合、归一化和AI原生查询才是真正的价值所在,金融数据本身极其杂乱,大规模清洗归一化的工程能力才是真正的准入壁垒。

笔记:点出了多数人忽略的核心壁垒,金融数据产品的核心竞争力不是AI模型,而是数据处理工程能力。

网友Edoardo
👍 1
追问视角

Coming from nuclear engineering into fintech with no traditional finance background, was that seen as a disadvantage early on or did the outsider perspective actually help you see the gap more clearly?

你从核工程跨界到金融科技,没有传统金融背景,早期这是劣势吗?还是说局外人视角反而让你更清晰地看到了市场空白?

笔记:提出了非常有价值的问题,跨界局外人视角往往更容易发现行业内习以为常的痛点。

网友carecrafted
👍 1
点出本质

At that scale, the real work usually shifts from building features to building systems around the product like onboarding, documentation, support workflows, data accuracy, compliance guardrails. Revenue becomes less about new features and more about operational execution.

达到这个规模之后,核心工作已经从开发新功能转向搭建周边系统:用户引导、文档、支持流程、数据准确性、合规防护,收入增长不再依赖新功能,而是靠运营落地能力。

笔记:道出了高门槛数据产品规模化后的核心增长逻辑,运营和合规比新功能更重要。

网友JarvisIdiogen
👍 1
深度认同

"Building data (DaaS) products takes a lot longer than software (SaaS) products." This is an underrated insight. Most founders underestimate data products because the complexity is invisible.

“做DaaS数据产品比普通SaaS耗时久得多”这个观点被严重低估,多数创始人低估了数据产品的隐形复杂度。

笔记:戳中了很多数据类创业者踩过的坑,前期要做好长期投入的心理准备。

网友Nanhe Gujral
👍 1
追问瓶颈

At a certain point, manual review, exception handling, and edge-case cleanup start becoming the real bottleneck — not model performance or distribution. Curious if there was a moment where operational load started increasing faster than revenue.

到了一定阶段,人工审核、异常处理、边缘案例清理才是真正的瓶颈,而不是模型性能或者分发,你有没有遇到过运营负载增长速度超过收入的阶段?

笔记:指出了AI数据产品规模化后的普遍痛点,提前搭建自动化异常处理体系非常重要。

网友Alexey Anshakov
👍 1
理性补充

But I'll push back on "trust your intuition." That advice only works if you've built enough pattern recognition. First-time founders trusting their gut often drive straight into a wall.

我对“相信直觉”这个观点有不同看法,这个建议只适用于已经积累了足够多行业经验、形成了模式识别能力的人,首次创业的人盲目信直觉很容易碰壁。

笔记:非常客观的补充,新手创业者不要盲目照搬成功创始人的经验,要结合自身情况判断。

网友hiktak
👍 1
经验共鸣

The “missing middle” you described is something I’ve felt too — either tools are overbuilt and inaccessible, or free but unusable. Bridging that gap with a clean, serious product is much harder than it looks.

你说的“中端空白”我深有体会,市场上要么是过度复杂价格昂贵的工具,要么是免费但完全没法用的产品,打造一款干净专业的中间层产品难度远超想象。

笔记:印证了这个市场空白的真实性,这类中端定位产品往往能获得非常高的用户忠诚度。

Localization Notes

  • 优先从A股/港股细分垂直行业金融数据切入,避开全量上市公司的高合规压力
  • 对接国内交易所公开披露的合规数据源,无需自行爬取非授权数据降低合规风险
  • 用国内云服务商阿里云/腾讯云替代海外Cloudflare,适配国内网络环境降低延迟
  • 先做小范围C端个人投资者验证产品价值,跑通PMF后再拓展B端金融机构数据授权
  • 接入国内主流大模型(通义千问/豆包/Kimi)完成异构金融数据归一化,降低清洗成本
  • 叠加投研告警、交易决策辅助功能,构建比纯数据产品更高的竞争壁垒