Asset-intensive organisations have invested heavily in enterprise asset management systems and condition monitoring infrastructure, yet most operational data remain underutilised. This presentation challenges the conventional centralised analytics approach and introduces edge federated machine learning as a transformative architecture for physical asset management.

Drawing from real-world implementations across Australian transport, ports, and heavy industry, this session demonstrates how edge computing combined with federated learning solves four critical barriers to intelligent maintenance:

• Isolated asset knowledge that prevents fleet-wide learning
• Data silos across facilities that block insight sharing
• Cloud analytics latency that misses split-second failure indicators
• Privacy concerns that lock away valuable operational data