Red Hat, NVIDIA & Palo Alto unveil AI-native telco stack

Red Hat has teamed up with NVIDIA and Palo Alto Networks on an architecture for telecommunications operators that combines cloud-native infrastructure, accelerated computing, and security controls for AI-driven network operations.

The work is positioned around AI shifting from an application-layer tool to a core part of how telecoms networks are operated. Operators are using AI for network optimisation, operational analytics, and automation. As AI moves closer to network infrastructure and distributed edge sites, the collaboration frames security as a foundational requirement.

The architecture centres on Red Hat OpenShift and Red Hat OpenShift AI. Together, they provide a shared environment for deploying and managing network functions and AI workloads across centralised data centres, edge locations, and remote network sites. The goal is to help operators keep operational processes consistent across different parts of the network.

Core Platform

OpenShift is Red Hat's Kubernetes-based platform for running containerised applications across hybrid and multi-cloud environments. In telecoms, Kubernetes platforms are used to host cloud-native network functions and supporting systems. OpenShift AI is positioned as the layer for building and operating AI workloads alongside those network functions.

Red Hat, NVIDIA, and Palo Alto Networks describe the combination as a common cloud-native platform for "AI-native telecommunications". The approach treats AI workloads as part of the same operational footprint as network workloads, rather than running them in separate environments.

NVIDIA's role is in the compute layer for AI workloads deployed across the network. The collaboration references NVIDIA RTX PRO Servers for AI acceleration in data centre and edge deployments, as well as the NVIDIA Aerial RAN Computer family of servers, marketed for radio access network and edge use cases.

In telecoms, AI inference and analytics often need to run closer to where data is generated, including cell sites, aggregation points, and edge data centres. These deployments raise operational and security challenges because operators need consistent management and controls across sites that vary in size, connectivity, and physical access.

Security Layer

Palo Alto Networks is providing security tools designed to run on OpenShift and OpenShift AI. The announcement highlights Prisma AIRS, a Palo Alto Networks product line positioned around AI security. It also references NVIDIA BlueField, which supports data processing and network acceleration within servers, and NVIDIA DOCA, a software framework for building applications that run on data processing units.

The approach combines centralised and distributed enforcement of security policies across the infrastructure stack, while also pointing to hardware-level functions for steering and acceleration. It is framed around the performance and latency requirements of telecoms workloads, particularly where AI supports real-time network operations.

Operators have been tightening security models as networks become more software-driven and distributed. The same shift is visible in enterprise IT, but telecoms networks also have strict availability requirements and specialised traffic flows. The collaboration draws on zero trust concepts, which treat internal network segments as untrusted and enforce verification and policy controls across systems and users.

The three companies are presenting the design as an integrated alternative to assembling separate components, linking it to consistent operations across network sites and aligned policy enforcement across core and edge infrastructure.

Telco Focus

The announcement cites AI-driven radio access network optimisation, predictive maintenance, energy efficiency, and automated security enforcement as areas where operators are exploring AI. Interest is rising as operators face growing network complexity, higher energy costs, and pressure to improve service quality.

RAN optimisation has become a prominent use case as mobile networks densify and operators deploy more spectrum bands and software-defined elements. Predictive maintenance uses telemetry and alarms to forecast failures or performance degradation. Energy efficiency is increasingly strategic as operators expand infrastructure footprints while also trying to reduce operating costs and emissions.

The architecture is also tied to longer-term network evolution. The announcement references "future 6G-era networks", pointing to early-stage work on next-generation mobile standards expected to incorporate more AI-driven optimisation and automation than earlier generations.

The collaboration was announced in the context of Mobile World Congress, the telecoms industry's annual gathering where vendors often unveil partnerships and product plans. Operators can use the event to discuss deployments and identify use cases for distributed AI workloads across telecoms infrastructure.