YC公司在AI热潮中取得成功的秘诀是什么
What's working for YC companies since the AI boom

原始链接: https://jamesin.substack.com/p/whats-working-for-yc-companies-since

这份分析通过考察2023年冬季、2023年夏季、2024年冬季和2024年夏季的Y Combinator (YC) 公司的A轮融资情况,探讨了自ChatGPT热潮以来哪些YC公司发展良好。在998家公司中,只有24家(2.4%)完成了A轮融资,但这其中存在偏差,因为许多公司尚未达到典型的种子轮到A轮融资的18个月时间线。 主要发现包括:业务自动化和运营工具占据主导地位,表明YC内部存在网络优势;API优先平台也表现强劲。令人惊讶的是,“AI赋能X”垂直领域较为狭窄,主要集中在法律和专利领域。值得注意的是,缺乏LLM工具、消费级、硬件和深度科技公司。种子轮顶级领投机构的存在似乎与A轮融资成功相关。 这项分析提出了一些问题,例如这些趋势与更广泛的市场相比如何,B2B SaaS与其他类别的成功率,以及YC估值上升对融资的影响。

Hacker News上的一篇讨论围绕YC支持的AI初创公司的表现展开。最初的帖子质疑了“AI包装器”初创公司的生存能力,认为它们的功能很快就会被OpenAI等大型公司复制。评论者对此展开了辩论,一些人强调了AI增强而非完全控制的潜力,另一些人则指出了用户界面和特定问题解决的重要性,而非通用AI能力。 讨论还涉及到缺乏面向消费者的AI初创公司获得A轮融资的情况,一些人将其归因于现有公司有效地整合了AI,消费者产品的推理成本高昂,以及强大的分销策略的重要性。硬件AI初创公司由于开发时间较长,难以获得融资。一些人建议AI初创公司应该专注于解决垂直领域的特定问题,而不是构建通用代理。很多人提到AI初创公司需要证明其商业价值和收入来源,但也有人指出一些公司从种子轮融资中获利,不需要A轮融资。
相关文章
  • (评论) 2025-05-11
  • YC:对初创公司的要求 2024-02-16
  • (评论) 2025-03-12
  • (评论) 2025-05-11
  • (评论) 2025-05-10

  • 原文

    The goal of this analysis was to understand what’s working for YC companies by looking at which ones have raised Series A rounds.

    I was really excited to take a look at the data for a number of reasons:

    1. We finally have 2 years of batch data since ChatGPT started the AI boom, which gives us some meat to analyze.

    2. YC is always interesting to dig into since it still has the cachet to attract top talent, and the batches are big enough to have a real population of companies that are representative of the market.

    3. Demo day is coming up, and we’ve been tracking the latest batch.

    4. First Round just announced Series A rounds for Reducto and David AI, and they have such an impressive hit rate - so I got curious whether there are actually tons of YC Series A rounds happening, or if First Round just crushes it (turns out they just crush it).

    Before diving into the findings, I’ll say the data is skewed and has a lot of limitations. For example, companies that raised large initial rounds (like Worldware's $30M from Spark) aren't going back to market anytime soon. We’re only seeing what gets announced on Crunchbase. I am only looking at the four batches directly after ChatGPT launched. The goal here is a quick snapshot of “what's working” since the market is moving so quickly, rather than a full breakdown of all YC batches or getting any answer to the question, “Does the YC model still work?”

    Final caveat - the topline number is: out of 998 companies from the first four batches since ChatGPT launched (Winter 2023, Summer 2023, Winter 2024, Summer 2024), only 24 have raised Series A rounds - just 2.4%.

    But this isn’t a great number to lean into as a takeaway. Most of these companies haven't had enough time to mature, since the typical Seed-to-Series A timeline is around 18 months, and only the Winter 2023 batch fits that window.

    So please take this post as surfacing directional trends rather than definitive statements.

    Ok! Let’s get into it. Here is what I found looking at Series A data since the Winter 2023 batch —

    Key takeaways from the 24 Series A companies include:

    • Business automation and operational tooling dominance. Probably the most surprising part of this analysis was how many of the winners were in internal business automation and operational platforms. The data looked different than the “what’s working in AI” analysis I did last year, which looked at Series A AI companies more broadly. There was less diversity here. This suggests the two obvious things:

      • Network advantages. The batches provide a built-in customer base for B2B operations and automation that can drive success (e.g, Deel, Brex). I don’t know who the customers are for the successful companies in this category, but I’m curious how many of them were accelerated by the YC batch advantage.

      • Technical talent. YC’s founder tends to be young, technical, and good at executing, which is an archetype that naturally gravitates towards automation infrastructure, developer tools, and general optimizations, perhaps rather than heavy vertically embedded software.

    • “AI for X” verticals are surprisingly narrow. Despite the hype around “AI for X” (e.g. AI for dentists), the only vertical AI categories that made it into the data are legal and patent-focused (e.g. Legora, Solve).

    • Platform/API-first success. 50% of successful companies are explicitly building platforms or APIs. This suggests YC’s breakout companies do lean into developer adoption and network effects for success and growth, and successful companies are not building one-off products.

    • There are notable absences from Series A success:

      • Zero LLM evaluation, observability, or tooling companies in the Series-A data.

      • Zero consumer, hardware, or deep tech companies in the Series-A data.

    • Top-tier lead investors matter. Most of the Seed rounds include top-tier lead investors — First Round Capital, General Catalyst, Uncork, Crat, Index, Soma, Greylock, Benchmark. Obviously, if one of these funds gives you a term sheet, it’s a no-brainer. However, we see party rounds at $2m on $20m (or $2m on $30m) with $200K left all the time pre-demo day, and this data suggests that having a reputable fund as a lead can be a better long-term setup for success.

    Along with this analysis, I built a simple website to interact with the companies and get a better visual of the categories.

    Check it out here:

    🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥

    https://yc-market-map-dot-chapter-one-340115.uc.r.appspot.com/

    🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥 🔥

    As mentioned above, the batches haven’t had much time to mature. Here is a table with the topline raise metrics:

    It’s worth mentioning that this data opened up a ton of questions that I’d love a chance to explore in the future:

    • How do these trends compare to the broader Seed-to-Series A market?

    • How much more likely is B2B SaaS to raise Series A versus other categories within YC?

    • As YC valuations have increased and terms have become less investor-friendly, are companies from recent batches struggling more with Series A fundraising compared to earlier batches? Or is the strategy working and aligned with the broader market?

    • Are there any patterns in founders’ backgrounds?

    Now, here are all the companies in the data broken down by category —

    Companies enabling businesses to optimize processes and build internal applications.

    • Artisan - Artisan automates your outbound with an all-in-one, AI-first platform powered by AI employees.

    • Reducto - Reducto is an AI-driven API that specializes in converting unstructured documents like PDFs and images into structured data.

    • Gumloop - Gumloop is a no-code platform to build and host AI-powered business automation.

    • Rollstack - Rollstack automatically creates and updates slide decks and documents.

    Platforms and integrations for back office operations (billing, taxes, accounting, contracts).

    • Capi Money - Capi Money is digital payment platform that provides budgeting tools, expense tracking, investment analysis services.

    • Numeral - Numeral puts sales tax on autopilot for leading e-commerce and SaaS businesses.

    • Pylon - Pylon is building the first customer support platform built for B2B companies.

    • Suger - Suger is a marketplace platform to manage product listings, offers, contracts, metering, and billing sevices.

    • Greenlite - Greenlite streamlines fintech and banking compliance using AI to assess company activities from financial statements and platform data.

    • Vooma - Vooma is a back-office automation platform that enables automated order entry for brokers and carriers.

    • Truewind - Truewind is an AI-powered accounting service company that offers bookkeeping and finance services.

    Tools enabling developer teams to focus on core products rather than building infrastructure.

    • Resend - Resend is an email API for developers.

    • Odigos - Odigos helps developers by generating distributed traces, metrics, and logs for any application.

    • ParadeDB - ParadeDB is a modern Elasticsearch alternative built on Postgres. Built for real-time, update-heavy workloads.

    • Fern - Start with an API spec. Generate SDKs in multiple languages and API documentation tailored to your brand.

    • Omnistrate - Omnistrate is a control-plane-as-a-service that transforms the user's docker image into a multi-cloud SaaS service.

    AI applications for specific industries.

    • Legora - Legora is a collaborative AI that empowers lawyers by enabling them to review documents, conduct research, and draft legal materials.

    • Solve Intelligence - Solve Intelligence is an artifical intelligence company that offers generative AI technology for writing patents.

    Audio and voice automation platforms.

    • David AI - The data layer for audio AI.

    • Bland AI - Bland AI provides a platform for realistic AI phone call automation..

    • Happy Robot - HappyRobot is a voice AI tool that automates phone operations used in the the logistics and fleet management sectors.

    • Flower Labs - Flower is a federated learning, analytics, and evaluation platform.

    • Persist AI - Persist AI is a biotechnology company that builds long-acting microsphere formulations.

    • Craftwork - Craftwork is a home service company that provides interior and exterior painting works.

    Thanks for reading! This was quick and high-level, but hopefully you found it interesting and useful. For any YC founders in this batch, good luck with Demo Day and the raise!

    I’m a General Partner at Chapter One, an early-stage venture fund that invests $500K - $2M checks into pre-seed and seed-stage startups.

    If you’re a founder building a company, please feel free to reach out on Twitter (@seidtweets) or Linkedin (https://www.linkedin.com/in/jamesin-seidel-5325b147/).

    联系我们 contact @ memedata.com