Meta has reportedly put all collaboration with AI recruitment firm Mercor on hold following a recent cyberattack that targeted the startup. The Menlo Park-based tech giant was among Mercor’s largest clients, relying on the company to hire subject matter experts who validate and perform quality analysis on outputs from large language models (LLMs). The breach is said to have compromised hundreds of gigabytes of sensitive data, prompting Mercor to launch an internal investigation into the incident.
According to a report by Wired, Meta’s pause on work with Mercor is indefinite. The publication cites unnamed sources familiar with the matter, who also noted that other major AI companies are reassessing their partnerships with the firm in the wake of the cyberattack. The move reflects growing caution within the AI industry, as companies evaluate the security and integrity of third-party partners that handle sensitive model validation work.
Mercor, founded in 2023, specializes in hiring domain experts to conduct quality checks on AI outputs. The startup has worked with several leading AI companies, including OpenAI and Anthropic, to ensure that large language models deliver accurate and reliable responses. Outsourcing this work allows AI firms to maintain model performance standards while continuously improving their systems based on expert feedback.
The company has attracted significant investment, having raised $350 million (roughly Rs. 3,257 crore) in a Series C funding round in October 2025, which valued Mercor at $10 billion (around Rs. 93,067 crore). Despite its rapid growth and high-profile partnerships, the recent security breach poses a serious challenge, highlighting the risks associated with handling large volumes of sensitive AI data and emphasizing the importance of cybersecurity in AI operations.