AI data labeling startup Handshake has announced the acquisition of Cleanlab, a company specializing in auditing data labels. This strategic move, confirmed by both companies to TechCrunch, aims to bolster Handshake's capabilities in providing high-quality data for foundational AI models. Handshake, initially launched in 2013 as a platform for hiring college graduates, expanded into human data labeling approximately a year ago. Cleanlab, founded in 2021, developed software designed to enhance the accuracy and quality of data generated by human labelers.
The acquisition is primarily an "acqui-hire," focused on integrating Cleanlab's talent into Handshake's research organization. Nine key Cleanlab employees, including co-founders Curtis Northcutt, Jonas Mueller, and Anish Athalye—all MIT computer science PhDs—will join Handshake. While the specific terms of the transaction remain undisclosed, acqui-hires can often prove surprisingly lucrative for founders, as previously reported by TechCrunch.
Cleanlab had successfully raised a total of $30 million from notable investors such as Menlo Ventures, TQ Ventures, Bain Capital Ventures, and Databricks Ventures. At its peak, the startup employed over 30 individuals.
Cleanlab's researchers are renowned for their expertise in developing algorithms capable of identifying incorrect data without requiring a second human review. This capability is crucial for Handshake, as the integration aims to significantly enhance the quality of data it provides 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? How high quality is that data?” Sahil Bhaiwala, Handshake’s Chief Strategy and Innovation Officer, told TechCrunch. “The Cleanlab team has been focusing on this problem for years.”
Curtis Northcutt, Cleanlab's CEO and a pioneer in automating data labeling auditing, confirmed that Cleanlab had received acquisition interest from several other AI data labeling companies. However, the startup ultimately chose Handshake. Northcutt explained that competitors like Mercor, Surge, and Scale AI frequently leverage Handshake's platform to source specialized human experts—such as doctors, lawyers, and scientists—for their data labeling projects.
“If you’re going to pick one, you should probably pick the source, not the middleman,” Northcutt remarked to TechCrunch.
Handshake, last valued at $3.3 billion in 2022, was projected to achieve an annualized revenue run rate (ARR) of $300 million by the end of 2025. The company is reportedly on track to reach an ARR in the "high hundreds of millions" this year. Handshake has already supplied data to eight leading AI labs, including industry giant OpenAI.
