Overview
Severity: CRITICAL | Affected: Coalesce AI | Category: breach
Coalesce AI, a leading generative AI firm, disclosed a major security breach attributed to a sophisticated state-sponsored threat actor. The attackers reportedly gained access to the company's core MLOps environment, exfiltrating the complete model weights for their flagship 'Fusion-5' large language model. In addition to this significant intellectual property theft, the breach also compromised a database containing millions of user prompts and fine-tuning datasets from enterprise clients, raising serious privacy concerns. The incident highlights the growing threat of corporate espionage targeting AI assets. Coalesce AI has initiated an internal investigation with law enforcement and cybersecurity firms, and is currently notifying affected customers. The full impact on the company's competitive standing and user trust is yet to be determined.