Snorkel AI Scores $85 Million for AI Data-Prep Platform

Chipbrain1

By: Mary Jander


Startup Snorkel AI, which shortcuts the manual labeling of data for machine learning (ML) and artificial intelligence (AI) applications, has scored $85 million in Series C funding on a valuation of over $1 billion – highlighting the momentum behind AI. The round brings Snorkel’s total raised to $135 million since its founding in 2019.

Snorkel offers a platform called Snorkel Flow that tackles a major challenge in the ML field – specifically, the need to label data in a manner that lets ML models consume information iteratively to solve problems and automate processes.

For instance, to “train” an ML model to detect cybersecurity threats, data identifying those threats, such as a pattern of code in malware, must be labeled before being fed iteratively to the system. Until recently, this kind of data preparation has been done manually by specialist firms.

But with demand rising for AI and ML apps in applications for government, manufacturing, finance, retail, and other areas, the need to reduce the slow and error-prone human element has grown. Snorkel offers a way to automatically label data items based on specific characteristics, sifting and honing the system over time to increase accuracy. Snorkel calls this a "programmatic" approach to labeling.

“Snorkel AI is addressing one of the most challenging and time-consuming bottlenecks in AI–the need for massive labeled datasets,” said Tony Kim, managing director at Snorkel investor BlackRock, in a press release. “With Snorkel Flow, enterprise customers have been able to develop and deploy AI applications in hours or days, versus months, leading to significant ROI.”

Over time, Snorkel AI, based in Palo Alto, Calif., has worked with Accenture (NYSE: ACN), Apple (Nasdaq: AAPL), Google (Nasdaq: GOOGL), IBM (NYSE: IBM), Intel (Nasdaq: INTC), and Microsoft (Nasdaq: MSFT), as well as DARPA, Stanford Medicine, Uber (NYSE: UBER), and an unnamed “top U.S. bank,” to hone its labeling system. The result is a platform endorsed by an impressive array of customers.

Snorkel’s Rise to the Surface

Snorkel was founded in 2019 by a cadre of Stanford Ph.D.s from the university’s AI Lab, including now-CEO Alex Ratner and Christopher Ré. The latter is a MacArthur Foundation “genius grant” winner who helped found SambaNova, an AI systems startup that won $676 million in funding this past April, as well as two other AI startups acquired by Apple – Lattice and Inductiv.

Christopher Ré, cofounder of Snorkel AI. Source: Snorkel AI

Indeed, Snorkel is part of a brain-trust of investors, key AI innovators, major U.S. research labs, and big customers. Tony Kim, for example, is on the board of advisors for SambaNova, which Ré cofounded. Wade Shen, who led ML- and AI-based programs at the Defense Advanced Research Projects Agency (DARPA), which also is a partner with Snorkel, recently joined SambaNova’s advisory board. BlackRock, Google Ventures, and Walden International are investors in both SambaNova and Snorkel. It doesn’t seem far-fetched to imagine future collaborations between SambaNova, which makes AI chips, and Snorkel.

AI and ML Business Booming

Snorkel competes with Scale AI, which got $325 million in a Series E funding round this past April on a valuation of more than $7 billion; and Labelbox, which garnered $40 million in February 2021. The rising tide of demand will no doubt swell as 5G emerges to produce more edge applications, including SASE, ensuring more market entrants. And the incestuous nature of the AI industry beckons collaborations and ecosystem permutations as well.

Snorkel’s round was led by BlackRock and Addition. New investors included Factory and Cooley. Existing investors who also joined included Greylock, GV, Lightspeed Venture Partners, Nepenthe Capital, Walden International, and Stone Bridge Ventures.