Date Published: July 27, 2026
Comments Due:
Email Comments to:
Author(s)
Yang Guo (NIST), Bennett Tomlinson (NIST)
Announcement
Building upon established principles from high-performance computing (HPC) threat analyses and security overlays, this publication delivers a thorough threat and security gap analysis for purpose-built AI infrastructure used in model training, inference, and applications. By contrasting AI data centers with traditional HPC systems across architecture, hardware, software stacks, workflows, and storage systems, the publication pinpoints critical security threats and outlines possible solutions to safeguard next-generation AI environments.
NOTE: A call for patent claims is included in this draft. For additional information, see the Information Technology Laboratory (ITL) Patent Policy--Inclusion of Patents in ITL Publications.
The architecture of high-performance computing (HPC) systems has significantly influenced the development of artificial intelligence (AI) data centers, which are purpose-built for model training, inference, and applications. This publication conducts a threat and security gap analysis of AI data centers and provides basic recommendations. It leverages prior work on HPC threat analysis, security posture, and the HPC security overlay to identify key differences between AI data centers and HPC systems in system architecture, hardware, software stacks, workflows, and data storage systems. The key threats facing AI data centers are then identified, and possible solutions are included.
The architecture of high-performance computing (HPC) systems has significantly influenced the development of artificial intelligence (AI) data centers, which are purpose-built for model training, inference, and applications. This publication conducts a threat and security gap analysis of AI data...
See full abstract
The architecture of high-performance computing (HPC) systems has significantly influenced the development of artificial intelligence (AI) data centers, which are purpose-built for model training, inference, and applications. This publication conducts a threat and security gap analysis of AI data centers and provides basic recommendations. It leverages prior work on HPC threat analysis, security posture, and the HPC security overlay to identify key differences between AI data centers and HPC systems in system architecture, hardware, software stacks, workflows, and data storage systems. The key threats facing AI data centers are then identified, and possible solutions are included.
Hide full abstract
Keywords
AI data center; AI data center security; AI data center computing environment; AI data center reference architecture; high-performance computing (HPC); HPC reference architecture; HPC security; security guideline
Control Families
None selected
