Snorkel AI是一家初创公司，致力于帮助AI实验室和企业构建训练数据集和模拟环境，近日以35亿美元的估值完成了3.5亿美元的E轮融资。

此轮由Insight Partners和S32领投，将这家成立七年的初创公司估值推高至近35亿美元，几乎是其17个月前完成1亿美元D轮融资时13亿美元估值的三倍。包括Addition、Lightspeed、Greylock、GV和富国银行在内的现有投资者也参与了本轮投资。

Snorkel最初提供数据标注自动化软件，但去年转向为客户提供完整的数据集，并将其称为“数据即服务”（data-as-a-service）。Snorkel不再单纯作为人类专家市场运营，而是采用混合方法，利用其软件和模型与领域专家共同合成生成数据。

Snorkel表示，其当前年化收入运行率已达3.75亿美元，在过去12个月内增长了18倍。这一增长得益于AI实验室对高端训练数据的不竭需求。

其他定位为AI数据实验室的数据公司也经历了类似的增长爆发。Mercor的总年化收入已攀升至20亿美元，Handshake今年早些时候突破了10亿美元里程碑，TechCrunch报道Micro1的收入规模已达到5亿美元。由于这些公司将约60%至70%的顶层收入直接支付给从事工作的领域专家，因此值得注意的是，它们的实际净年收入远低于这些 headline 总营收数字。

Snorkel表示，鉴于其销售的是强化学习（RL）环境和完整数据集而非人力劳动，支付给人类专家的款项计入其销货成本，而非产生头条新闻的年化收入数字。

在联合创始人兼首席执行官Alex Ratner及其团队在斯坦福AI实验室进行四年研究后，Snorkel于2019年正式商业化推出。

Snorkel AI, a startup that helps AI labs and corporations build training data sets and simulated environments, has raised a $350 million Series E at a $3.5 billion valuation.

The new round, which was led by Insight Partners and S32, valued the seven-year-old startup at nearly triple the $1.3 billion valuation it garnered when it raised $100 million in a Series D 17 months ago. Existing investors, including Addition, Lightspeed, Greylock, GV, and Wells Fargo, also participated in the round.

While Snorkel originally provided software for data labeling automation , it shifted last year to providing customers with completed data sets, an offering it calls data-as-a-service. Rather than operating purely as a human expert marketplace, Snorkel relies on a hybrid approach, using its software and models to generate data synthetically alongside subject matter experts.

Snorkel says its current annualized revenue run-rate now stands at $375 million, an 18-fold increase over the last 12 months. That growth is fueled by AI labs’ insatiable appetite for high-end training data.

Other data companies positioning themselves as AI data labs have seen a similar explosion in growth. Mercor’s gross annualized revenue has climbed to $2 billion , Handshake hit the $1 billion milestone earlier this year, and TechCrunch reported that Micro1 has scaled to $500 million . Since these companies pay out roughly 60% to 70% of their top-line income directly to the domain specialists doing the work, it’s important to note that their actual net annual revenue is substantially lower than those headline gross figures.

Given that Snorkel sells reinforcement learning (RL) environments and complete datasets rather than human labor, payments to its human experts are accounted for in its cost of goods sold rather than headline-generating annualized revenue numbers, according to the company.

Snorkel launched commercially in 2019 following four years of research by co-founder and CEO Alex Ratner and his team at a Stanford AI lab.