Red Hat Outlines AI-RAN Roadmap
RCR Wireless News, Friday, July 17th, 2026
Red Hat maps a three-phase AI-RAN evolution through 2030, urging targeted GPU deployments over blanket acceleration.
Red Hat's Shujaur Mufti laid out a phased AI-RAN evolution: AI for RAN through roughly 2027, AI and RAN sharing infrastructure from 2027 to 2030, and AI on RAN enabling native AI applications after 2030.
Operators favor AI-RAN because it yields measurable operational benefits without major radio network upgrades, including lower operating costs, better energy and spectral efficiency, improved performance and fault detection.
Mufti cautioned against assuming universal GPU adoption, recommending operators start where the business case is strongest, typically edge AI inferencing before adding RAN workloads.
Red Hat is working with SoftBank, Fujitsu and NVIDIA on early GPU-accelerated deployments that have run Layer 1 and Layer 2 functions without real-time kernels.