AuthMind

Cybersecurity · Identity Security · AI Agents

Acquisition

IBM acquires AuthMind to secure AI agent credentials inside Vault

The deal fills the gap after credentials are issued, when agents actually use their access across systems.

• IBM announced on November 12, 2026 that it acquired identity security startup AuthMind, folding it into the HashiCorp Vault portfolio; deal terms were not disclosed.
• Founded in 2020 and led by CEO Shlomi Yanai, AuthMind built an identity activity graph that tracks how human, machine, and AI agent identities actually use credentials, not just when they're granted.
• AuthMind analyzes network traffic to discover shadow or unsanctioned AI agents and flag unusual access patterns, excessive permissions, and policy violations after credentials leave Vault.
• The companies were already linked: IBM Ventures backed AuthMind's 2023 seed round, and AuthMind had integrated with IBM Identity Protection before the acquisition.
• AuthMind had raised $27.8 million total, including a $19.3 million seed round led by Cheyenne Ventures in April 2025, and employed 62 people at the time of the deal.
• IBM VP Suja Viswesan called unauthorized AI activity 'rogue agents,' describing the AuthMind integration as a way to stop agents from acting outside their authorized scope.
• IBM exited cybersecurity assets in 2024 by selling QRadar to Palo Alto Networks for roughly $1.1 billion, making this buy a notable re-entry into the sector just as rivals like Palo Alto spend far more (its $25 billion CyberArk deal) on identity security.
• As enterprises deploy autonomous AI agents that act continuously and delegate tasks without human oversight, monitoring what credentials actually do after issuance is becoming as critical as issuing them in the first place.