The discussion frames key industry choices about where to invest — capex on GPU-equipped sites versus centralized/near-edge cloud — and whether AI will shift telcos from cost-cutting to new revenue streams, influencing vendor strategies and future network architecture.
Nokia’s Sunday event in Barcelona proved underwhelming, offering no major new customer deals or breakthrough announcements despite last year’s $1 billion Nvidia partnership to pursue an AI-driven RAN strategy. Speakers reiterated the dual promises of AI for RAN—cost and spectral-efficiency gains—and the potential to repurpose GPU-enabled infrastructure for edge AI inference aimed at enterprise use cases. Panelists debated practical needs for ultra-low latency, the number and placement of edge facilities, and whether telcos can realistically monetize compute at cell sites versus relying on near-edge cloud providers. Energy efficiency and whether GPUs are the optimal network solution remained open points of contention.
Comments
Want to join the conversation?
Loading comments...