Databricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.
Databricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.
Databricks 原计划融资 10 亿美元，投资者却想投 150 亿美元，最终以 1900 亿美元估值敲定 50 亿美元融资。
There’s a funny kind of game that the latest of late-stage startups must play when raising money. They often have to sell more shares than they want or risk offending some of their existing VCs. This scenario recently played out with AI big-data company Databricks and its latest $5 billion raise announced Thursday, co-founder and CEO Ali Ghodsi (pictured above) told TechCrunch. 在融资时，处于后期阶段的初创公司往往不得不玩一种有趣的“游戏”。它们经常被迫出售比预期更多的股份，否则就有可能得罪现有的风险投资人。据联合创始人兼首席执行官 Ali Ghodsi（见上图）向 TechCrunch 透露，人工智能大数据公司 Databricks 最近在周四宣布的 50 亿美元融资中就经历了这一局面。
“We wanted to raise $1 billion, but then The Information printed this article saying that Databricks is doing a big fundraise. They did that in the middle of our conference. We were heads down with our conference, and we were not actually at all focused on fundraising,” Ghodsi recalled, referring to a conference that took place in June. “As soon as that article went out, there was a long line of investors that started calling. My phone blew up. It was like the worst timing for us because we were busy with our conference,” he said. “我们原本只想融资 10 亿美元，但《The Information》刊登了一篇报道，称 Databricks 正在进行大规模融资。他们是在我们举办大会期间发布这篇报道的。当时我们正全身心投入大会，根本没考虑融资的事，”Ghodsi 回忆起六月份的那场会议时说道。“那篇文章一出，排着队的投资者就开始打电话。我的手机被打爆了。这对我们来说时机简直糟透了，因为我们正忙于会议。”
It was an enviable problem that turned the news report into a self-fulfilling prophecy. “The interest level was just insane. Just from this select group of investors that we looked at, there was $15 billion of interest,” he said. When there’s that much desire to get into a deal, telling some long-term backers no is a recipe for hard feelings. Databricks decided to issue more stock, and in July, sent out a press release announcing it had closed its new round at a $188 billion valuation. (The company didn’t disclose at the time how much it had raised.) 这是一个令人羡慕的“烦恼”，这篇新闻报道最终成了一个自我实现的预言。“投资意愿简直疯狂。仅从我们接触的这部分精选投资者来看，意向投资额就达到了 150 亿美元，”他说。当有如此强烈的入局意愿时，拒绝一些长期支持者只会导致不愉快。Databricks 决定增发股票，并于 7 月发布新闻稿，宣布以 1880 亿美元的估值完成了新一轮融资。（当时公司并未披露具体融资金额。）
On Thursday, Databricks shared it raised $5 billion from a paragraph worth of VCs that it let in on the deal and that its valuation pushed higher to a nice round $190 billion. The $5 billion round was led by Coatue and several others, including Blackstone, MGX, various accounts associated with various arms of T. Rowe Price, and new investor Sixth Street Growth. (Sixth Street is the firm founded by former Goldman Sachs chief investment officer Alan Waxman.) About two dozen VCs were named as participants. 周四，Databricks 分享了最新进展：公司从一长串风险投资机构中筹集了 50 亿美元，估值也进一步推高至 1900 亿美元的整数关口。这轮 50 亿美元的融资由 Coatue 领投，其他参与方包括 Blackstone、MGX、与 T. Rowe Price 各部门相关的多个账户，以及新投资者 Sixth Street Growth。（Sixth Street 是由前高盛首席投资官 Alan Waxman 创立的公司。）约有两打风险投资机构被列为参与者。
Why were they all so eager? Databricks seems like a sure bet. Ghodsi said his company has hit $7 billion of annualized run rate revenue, which is currently growing at 80% and is cash-flow positive. Its core product, a cloud data warehouse, is $1.5 billion of that run rate, and still growing at 100% year-over-year, he said. Plus, Databricks has the magic AI pixie dust. Its database for agents, Lakebase, launched in June, 2025, and has hit $100 million revenue run rate. Its AI chatbot tool Genie, that can do business analysis on the spot, “is insanely popular,” he said. 为什么他们都如此渴望入局？Databricks 看起来是一个稳赚不赔的赌注。Ghodsi 表示，公司年化营收运行率已达到 70 亿美元，目前增长率为 80%，且现金流为正。他说，其核心产品云数据仓库贡献了 15 亿美元的运行率，且同比增速仍保持在 100%。此外，Databricks 还拥有人工智能的“魔法粉末”。其于 2025 年 6 月推出的智能体数据库 Lakebase，年化营收运行率已达到 1 亿美元。他表示，其能够进行即时业务分析的 AI 聊天机器人工具 Genie “极其受欢迎”。
So, if the business is doing so well, why raise more capital? The company had already raised $20 billion over the past 20 months. AI is expensive, Ghodsi said. Databricks has multibillion-dollar cloud commitments with all three of the major hyperscalers. On top of that, “AI research is very expensive,” he said, adding that the company has an AI research team of 100 people, a highly competitive area. Plus, Databricks is shopping. “We do a lot of M&A.” 既然业务表现如此出色，为什么要筹集更多资金？该公司在过去 20 个月里已经筹集了 200 亿美元。Ghodsi 说，AI 很昂贵。Databricks 与三大超大规模云服务商都有数十亿美元的云服务承诺。除此之外，“AI 研究非常昂贵，”他补充道，公司拥有一个 100 人的 AI 研究团队，这是一个竞争极其激烈的领域。此外，Databricks 还在“买买买”。“我们进行了大量的并购。”
Ghodsi said, referencing an acquisition the company announced this week of Electric, the company that makes the lightweight Postgres database PGlite, a means for agents to spin up databases (terms undisclosed). In June, it bought AI cybersecurity company Panther; in March, it bought two startups. There was a time when a $1 billion round was considered a massive and difficult raise. In this age of AI spending, where startups are raising $1 billion for a seed/Series A right out of the gate, that amount is now a pittance. Ghodsi 提到公司本周宣布收购 Electric，该公司开发了轻量级 Postgres 数据库 PGlite，这是一种让智能体快速启动数据库的工具（条款未披露）。6 月，它收购了 AI 网络安全公司 Panther；3 月，它收购了两家初创公司。曾几何时，10 亿美元的融资被认为是规模巨大且难以完成的。但在如今这个 AI 烧钱的时代，初创公司在种子轮或 A 轮就能筹集 10 亿美元，这个数字现在看来只是九牛一毛。
Still, Databricks’ private fundraising, instead of going public, has become something of a meme among the Valley. When it announced this round last month, people joked online that it has raised so many, it was running out of letters of the alphabet. Ghodsi told CNBC that he still wants to take the company public one day. With such a giant roster of investors who will want to cash out one day, how can he promise anything else? But today, he wants to focus on investing in AI, he said. Given the expenses involved in that, perhaps doing so out of the public eye is a wise idea. Plus, when he can command an instant $15 billion of interest, and on his own terms, what’s the rush? 尽管如此，Databricks 选择私下融资而非上市，已成为硅谷的一个“梗”。当上个月宣布这一轮融资时，人们在网上开玩笑说，它筹集的次数太多，以至于快要把字母表用完了。Ghodsi 告诉 CNBC，他仍然希望有一天能让公司上市。面对如此庞大的投资者阵容，他们终有一天需要套现，他还能承诺什么呢？但他表示，今天他只想专注于 AI 投资。考虑到其中涉及的巨额开支，或许在公众视野之外进行这些操作是明智之举。况且，当他能随心所欲地获得 150 亿美元的投资意向时，又何必急于一时呢？