2 Minutes
Meta has quietly put the brakes on its work with Mercor after a security breach at the fast-growing AI training startup, a move that underscores just how fragile the supply chain behind modern artificial intelligence can be.
According to a source familiar with the matter, the company is now reviewing what happened and how much exposure it may have had. Wired first reported that Meta had paused all work with Mercor, and Business Insider later confirmed the development through a separate source.
Mercor is no small player. The startup, which was valued at $10 billion in a funding round in October, helps major technology companies train AI systems by connecting them with thousands of human contractors and subject-matter experts. That kind of human-in-the-loop model has become a critical part of building and refining large language models, where raw computing power still needs a human touch.
A breach with wider ripple effects
Mercor confirmed the incident on Friday, saying it had been affected by a supply chain attack involving LiteLLM, an open source project used by thousands of companies. The startup said it identified the issue, moved quickly to contain it, and brought in third-party forensic experts to help investigate.
The incident is a reminder that even companies powering the next wave of AI can be vulnerable to the same cyber risks that plague the rest of the tech industry.
Meta declined to comment on the pause, and it remains unclear how long the review will take or whether the relationship between the two companies will resume in full. For now, the message is clear: in AI, trust is becoming as valuable as technical capability.
Leave a Comment