• New addendum covers GPU density, liquid cooling and AI-ready benchmarks.
  • Google and major operators join supply-chain quality standard push.

What happened

The Telecommunications Industry Association (TIA) is expanding beyond its telecoms roots with new standards aimed at AI-driven data centres. The group detailed an addendum to its ANSI/TIA-942 standard, designed specifically for high-density, AI-focused environments.

The revised 942 framework will guide operators deploying GPU-heavy clusters, liquid cooling systems and high-performance compute infrastructure. It targets hyperscale and enterprise facilities facing rising rack densities and thermal loads driven by AI workloads.

Alongside this, TIA is developing the DCE 9000 series, a quality management standard focused on data centre infrastructure supply chains. The initiative includes participation from major industry players such as Google, alongside vendors and operators. It aims to introduce consistent processes for design, deployment and operational quality across the ecosystem.

TIA is also extending its certification programmes to align with these updates, enabling facilities to be assessed against AI-ready benchmarks. The organisation further highlighted the growing importance of edge data centres, signalling a need for standards that support distributed, lower-latency deployments beyond core campuses.

Why it’s important

TIA’s initiatives signal a shift in how data centres are defined. What began as telecoms infrastructure is now evolving into core AI infrastructure — built for high-density compute, advanced cooling and integrated systems. Supply chain standards and edge considerations extend this transition from centralised campuses to distributed environments.

Industry organisations are no longer just adapting to AI. They are defining it — marking a turning point as data centres shift from communications infrastructure to foundational platforms for the AI economy.

Also read: Africa Data Centres boosts SA connectivity with Oni-Tel

Also read: Anthropic launches Project Glasswing with Nvidia and Cisco for AI security