Meta has stopped working with Mercor, an AI recruitment company.
A cyberattack hit Mercor last week.
The attack caused a major data breach.
Hackers claimed to steal hundreds of gigabytes of data from the firm.
Mercor is now investigating the incident.
Meta, based in Menlo Park, was one of Mercor’s biggest clients.
The startup hires experts to check and improve output from large language models.
According to a Wired report, Meta has paused all work with Mercor indefinitely.
Two sources familiar with the matter shared this information.
Other major AI companies have also started reviewing their ties with Mercor.
Mercor started in 2023.
It partners with firms like Anthropic and OpenAI.
These companies outsource quality checks to validate and enhance their AI models.
In October 2025, Mercor raised $350 million in Series C funding.
Its valuation then reached $10 billion.
Mercor released a statement about the attack.
The company said it was affected by a supply chain attack involving LiteLLM.
It is working with top forensics experts for a full investigation.
Mercor plans to update its customers and contractors soon.
It will use all necessary resources to fix the issue quickly.
Meanwhile, the hacking group LAPSUS$ took responsibility for the breach.
They claimed to have stolen nearly 4TB of data.
This included databases, source code, and bucket data.
The group is reportedly auctioning the stolen information.
However, no one has verified these claims yet.
Cyber experts continue to watch the situation closely.
This event highlights the growing risks in the AI supply chain.
Companies are now paying more attention to partner security.
As a result, AI firms may tighten their vendor checks in the future.