Conference

From Edge to Everywhere: Building the Distributed AI Network

Pass Required to Access Physical

Date

Tue, 2 Mar

Time

9:30 am - 10:30 am CET

Location

Turing Stage, Hall 6


Loading…

Session Description

AI's centre of gravity is shifting. While models are still trained in the cloud, where compute is abundant, inference is increasingly moving closer to where data is created, running in parallel across devices, edge sites and local servers to improve efficiency, scalability and responsiveness. Distributed AI, a concept first explored in the mid-1970s, is now becoming an operational reality across telecoms and technology enterprises. 

For telcos, this creates a rare opportunity. Their footprint of network assets, edge sites and data centres already spans the locations where AI workloads need to run. Emerging approaches such as Distributed Machine Learning and Federated Split Learning allow models to be trained collaboratively across decentralised nodes without pooling raw data, combining the scalability of federated learning with the privacy and efficiency of split learning, which is critical for regulated and privacy-sensitive industries. 

This session explores whether operators can turn their distributed infrastructure into a new AI compute layer, what it will take to orchestrate inference seamlessly across cloud and edge, and where the real-world use cases and business models are emerging.


Similar Sessions