AI factories and identity controls: are your guardrails ready?

TL;DR: AI factories move AI development on-premises or into hybrid data centres, giving organisations more control but also shifting responsibility for access management, auditing, and privileged control onto internal teams, according to Delinea. The critical gap is that high-performance AI environments can amplify unmanaged service identities, shadow AI, and over-privileged access faster than standard IAM processes can keep up.

NHIMG editorial — based on content published by Delinea: Secure AI factories with NIST HPC guidelines, NVIDIA architecture, and Delinea controls

By the numbers:

• Only 5.7% of organisations have full visibility into their service accounts.
• 97% of NHIs carry excessive privileges, increasing unauthorised access and broadening the attack surface.

Questions worth separating out

Q: How should security teams govern service accounts in AI factories?

A: Security teams should treat service accounts in AI factories as high-value non-human identities with clear owners, scope, and expiry dates.

Q: Why do AI factories increase the risk of privilege creep?

A: AI factories increase privilege creep because performance-sensitive clusters encourage broad operational access, temporary exceptions, and account reuse across jobs and tools.

Q: What breaks when AI workloads run outside zone-based controls?

A: When AI workloads run outside zone-based controls, access becomes too broad to audit and too easy to reuse.

Practitioner guidance

• Inventory every machine identity in the AI factory Build a register for service accounts, scheduler identities, AI microservices, and control-plane credentials.
• Enforce zone-specific privilege boundaries Map access, management, compute, and storage zones to separate roles and deny direct cross-zone access unless the request is explicitly approved and logged.
• Rotate keys and principals for automated services Automate credential rotation for Kerberos principals, keytabs, and service credentials used by AI pipelines.

What's in the full article

Delinea's full blog covers the implementation detail this post intentionally leaves at the architectural level:

• The NIST SP 800-234 control mapping for AI factories, including account management, authentication, auditing, and user-installed software controls.
• The zone-by-zone operating model for access, management, compute, and storage boundaries in an AI factory.
• How Delinea applies Kerberos, AD integration, and privileged session management across Linux-based AI clusters.
• The specific session recording, logging, and service account automation features used to support administration and forensic review.

👉 Read Delinea's analysis of NIST controls and identity security for AI factories →

AI factories and identity controls: are your guardrails ready?

Explore further

View Full Forum →  |  NHI Foundation Course →