AI data labeler Handshake acquires Cleanlab amid strong acquisition interest

AI data-labelling company Handshake has acquired data–labelling startup Cleanlab, the two companies confirmed. Cleanlab had drawn acquisition interest from multiple firms in the AI data-labelling space before the deal.

Handshake was founded in 2013 as a platform focused on hiring college graduates and expanded into human data labelling roughly a year ago, targeting companies that build large foundational AI models. Cleanlab, launched in 2021, develops software designed to improve the accuracy and quality of data generated by humanlabellerss.

The transaction is primarily an acqui-hire, according to the companies, with Handshake adding nine key Cleanlab employees to its research organisation. Among them are Cleanlab’s three co-founders — Curtis Northcutt, Jonas Mueller, and Anish Athalye — all of whom earned PhDs in computer science from MIT. Financial terms of the acquisition were not disclosed, though acqui-hires can sometimes result in significant payouts for founders.

Cleanlab has raised a total of $30 million in venture funding from investors including Menlo Ventures, TQ Ventures, Bain Capital Ventures, and Databricks Ventures. At its peak, the startup employed more than 30 people.

The Cleanlab team specialises in developing algorithms that automatically identify incorrect or low-quality data without requiring a second human reviewer. Handshake plans to use this expertise to improve the quality of the labelled data it supplies to AI labs.

“We have an in-house research team that thinks a lot about where our models are weak, what data should we be producing, and how high quality that data is,” said Sahil Bhaiwala, Handshake’s chief strategy and innovation officer. “The Cleanlab team has been focused on this problem for years.”

Northcutt said Cleanlab had received acquisition offers from other AI data-labelling companies but ultimately chose Handshake because many competing firms already rely on Handshake’s platform to source specialised human experts, including doctors, lawyers, and scientists, for labelling projects.

“If you’re going to pick one, you should probably pick the source, not the middleman,” Northcutt said.

Handshake was last valued at $3.3 billion in 2022 and was projected to close 2025 with anannualisedd revenue run rate of $300 million. The company is reportedly on pace to reach an ARR in the high hundreds of millions this year and has provided data to eight leading AI labs, including OpenAI.