Inside the 40-Minute Meltdown: How a Tiny Supply Chain Flaw Exposed Mercor’s AI Empire
A fleeting software compromise ripples through the tech world, as hackers claim a massive data heist from a $10B AI unicorn.
It only took 40 minutes. That’s how long cybercriminals needed to slip through the digital defenses of Mercor, a leading AI recruitment firm, and potentially walk away with a staggering 4 terabytes of sensitive data. The breach - rooted in a poisoned software update - now threatens to upend the trust in open-source tools powering the world’s most advanced AI systems.
Fast Facts
- Mercor, valued at $10 billion, confirmed a breach after hackers exploited a compromised open-source tool.
- Attackers published two malicious versions of the LiteLLM package, available for just 40 minutes in March 2026.
- Hacker groups TeamPCP and Lapsus$ are implicated, with the latter claiming to have stolen 4TB of data.
- Compromised data allegedly includes candidate profiles, source code, API keys, and sensitive interviews.
- The attack is part of a larger supply chain compromise affecting thousands of organizations worldwide.
The Anatomy of a Digital Heist
Late March 2026, a seemingly routine update to LiteLLM - a popular open-source tool connecting AI models - set off a global chain reaction. Attackers, allegedly from the TeamPCP group, managed to upload two tainted versions of LiteLLM to the Python Package Index (PyPI). For just 40 minutes, these malicious packages were live. But with millions of downloads per day and widespread use in automated cloud environments, that brief window was all it took.
Mercor, which connects tech giants like OpenAI and Anthropic with domain experts, was among the thousands caught in the crossfire. The attackers exploited compromised maintainer credentials, a tactic that has become alarmingly common in recent supply chain attacks. The breach reportedly traces back even further, to an earlier compromise involving the Trivy tool, which leaked sensitive access tokens - digital keys that unlock downstream systems.
As Mercor scrambled to contain the fallout, the notorious Lapsus$ extortion group surfaced, claiming to possess 4TB of Mercor’s most sensitive data. Their leak site listing boasted everything from candidate profiles and employer records to source code, API secrets, and even video interviews between AI systems and contractors. While Mercor acknowledged the breach, it has not verified the extent or authenticity of Lapsus$’ claims.
Security experts believe there may be a link between TeamPCP and Lapsus$, but the nature of their relationship remains murky. Notably, the Mercor auction was later removed from Lapsus$'s site - fueling speculation about behind-the-scenes negotiations or a successful data sale, though nothing is confirmed.
This incident underscores a chilling reality: even a fleeting vulnerability in trusted open-source software can cascade into a global crisis, endangering the data of thousands of organizations in minutes.
Aftershocks and Lessons
As Mercor works with forensic experts to untangle the breach, the wider tech community faces a reckoning. The very openness that powers modern AI development can also open doors for attackers. For now, the only certainty is that the battle for the software supply chain’s integrity is just beginning - and no one is immune.
WIKICROOK
- Supply Chain Attack: A supply chain attack is a cyberattack that compromises trusted software or hardware providers, spreading malware or vulnerabilities to many organizations at once.
- Open: 'Open' means software or code is publicly available, allowing anyone to access, modify, or use it - including for malicious purposes.
- PyPI (Python Package Index): PyPI is the official online repository for Python packages, letting developers upload, share, and download reusable code libraries and tools.
- API Key: An API key is a unique code that lets programs access data or services. If not properly secured, it can pose a cybersecurity risk.
- CI/CD Pipeline: A CI/CD pipeline automates code testing and deployment, enabling developers to deliver software updates quickly, reliably, and with fewer errors.