Scale AI, a company that helps tech companies prepare data to train AI models, has filed a lawsuit against one of its former sales employees and its rival, Mercor. The suit alleges the employee, Eugene Ling, who was hired by Mercor, stole more than one hundred confidential documents containing Scale’s customer strategies and other proprietary information.
Scale is suing Mercor for misappropriation of trade secrets and is suing Ling for breach of contract. The lawsuit further claims the employee was attempting to pitch Mercor to one of Scale’s largest customers, referred to as Customer A, before he had officially left the company.
In response, Mercor co-founder Surya Midha denied that his company used any data from Scale, though he admitted Ling may have possessed some documents. Midha stated that while Mercor has hired many former Scale employees, it has no interest in Scale’s trade secrets and is intentionally running its business differently. He informed TechCrunch that Ling had old documents in a personal Google Drive which Mercor has never accessed and is now investigating. Midha also said Mercor reached out to Scale six days prior offering to have Ling destroy the files or reach a resolution and is currently awaiting a response.
Scale alleges these documents contained specific data that would enable Mercor to serve Customer A and several other key clients. The company wanted Mercor to provide a full list of the files in the drive and to prevent Ling from working with Customer A. The suit alleges Mercor refused these requests. Ling did not immediately respond to a request for comment.
There are few clues in the suit about the identity of Customer A, though it states that winning this customer away would be a contract worth millions of dollars to Mercor.
This lawsuit demonstrates that Scale is sufficiently concerned about the threat from Mercor to pursue legal action. As previously reported, even with Meta’s multibillion-dollar investment into Scale, the core unit within Meta tasked with building AI superintelligence continues to use Mercor and other data training providers.
Mercor is gaining prominence in the large language model training arena because it is known for hiring content specialists, often PhDs, to train data within their fields of expertise. In June, Scale announced that Meta was investing $14.3 billion for a 49% stake in the company and was hiring away its founder. Shortly after that, several of Scale’s largest data customers, who are competitors to Meta, reportedly cut ties with the company.