Building Profitable AI-Native Networks From 5G-Advanced to 6G

Why It Matters

By embedding GPU‑powered AI at the edge, telcos can extract more value from scarce spectrum and turn idle network compute into a lucrative AI service, fundamentally changing revenue and efficiency dynamics.

Key Takeaways

• Nvidia invests $1B with Nokia to embed GPU at edge.
• Software‑defined, programmable base stations enable AI workloads on RAN.
• Three AI‑RAN models: AI on RAN, AI for RAN, AI + RAN.
• Edge AI improves spectral efficiency and reduces operational costs.
• Multi‑purpose compute can monetize idle capacity via token services.

Summary

The panel on Telecom TV examined how AI will reshape radio access networks from 5G‑Advanced to 6G, spotlighting Nvidia’s $1 billion partnership with Nokia to place GPU‑accelerated compute directly at the cell‑site edge. By moving inference to the base‑station, operators can run real‑time AI, boost spectral efficiency and transition from a connectivity‑only model to an "intelligence‑connected" network.

Speakers broke the AI‑RAN landscape into three distinct layers: AI on RAN (applications such as vision, robotics, or token‑based services running atop the radio), AI for RAN (machine‑learning algorithms that improve beamforming, channel estimation and overall spectrum utilization), and AI + RAN (a shared, software‑defined compute platform that can dynamically allocate resources between radio functions and AI workloads). The discussion emphasized that programmable, software‑defined base stations are the prerequisite for this multi‑purpose architecture.

Nvidia’s vice‑president highlighted a live demo where drone‑detection video feeds were processed at the edge, illustrating the low‑latency, secure inference possible on the same hardware that runs Nokia’s RAN stack. Nokia’s AI‑for‑RAN team cited parallel GPU compute and new ML models that can raise gigabits‑per‑dollar and cut operational overhead, while Nvidia stressed the market potential of “token‑as‑a‑service” running on under‑utilized compute.

The convergence promises operators a dual upside: dramatically higher spectrum efficiency and a new revenue stream from AI services, pushing compute utilization from roughly 30 % to 80‑90 %. As networks become fully programmable and AI‑native, they will act as platforms for cross‑industry innovation, accelerating the rollout of AI‑driven cities and reshaping the telecom business model.

Original Description

This TelecomTV webinar brought together experts from Nokia and NVIDIA to explore how AI-RAN is reshaping mobile networks. GPU-accelerated AI-RAN enables the transition from 5G to AI-native 6G, helping mobile network operators benefit from the fusion of AI and RAN. It scales up accelerated AI computing for highly efficient 5G-Advanced and AI-native 6G (AI-for-RAN), leveraging computing and infrastructure synergies (AI-and-RAN) and unlocking new monetisation opportunities (AI-on-RAN).

The discussion highlighted what telecom providers need in order to capture the business opportunity unlocked by AI-RAN and explored the growing importance of programmable networks for enabling the connection of network capabilities with enterprise applications, creating opportunities for cross-industry innovation.

For more information please visit nokia.com - https://link.telecomtv.com/nokia-ai-ran