At first, AI was very much a cloud story. Everyone wanted to train new LLMs and models. And that was best done in centralised hubs due to cost and infrastructure needs.
Now, however, we have evolved into an Age of Inference which is rapidly moving AI to the edge as customers seek reduced latency and lower cloud egress costs.
The opportunity for telcos is rapidly shifting.
Customer adoption has also changed the shape of capacity demand for telcos. The leaders are finding that AI introduces unpredictable traffic patterns and massive scalability requirements into the network. That is putting pressure on telco leaders to ramp up their network investments to ensure they are offering high-capacity and highly meshed networks that can adapt to changing demand curves while supporting AI-driven orchestration.