The Connectivity Problem Holding Back Edge AI

Michael De Nil is the CEO and cofounder at Morse Micro—reinventing Wi-Fi for IoT.

​For years, connected infrastructure was largely about monitoring: sensors collected data, systems sent alerts and operators responded after the fact.

That model is changing. Cameras, meters, energy systems, access control devices and industrial sensors are becoming intelligent endpoints that support real-time decisions at the edge.

This shift creates a new challenge. Edge AI cannot operate everywhere if connectivity cannot reach everywhere. If connectivity is unreliable, too power-hungry, too short-range or too complex to scale, the intelligence built into those systems cannot be fully realized.

Today, smart meters, perimeter cameras, HVAC systems, EV chargers, solar panels, battery storage and water-monitoring systems are becoming part of a larger operational network that depends on reliable, persistent communication. As more systems move “behind the meter,” connectivity is essential supporting infrastructure.

From my work across wireless semiconductors and next-generation connectivity, I’ve seen this shift firsthand as more devices move from passive sensing to real-time intelligence at the edge. The pattern is consistent across buildings, energy systems, security and industrial environments: connectivity is becoming a core design constraint.​

This shift exposes several challenges that traditional wireless architectures were never originally designed to solve.

1. Coverage was designed around people, not intelligent infrastructure.

Historically, wireless coverage priorities focused on living rooms, office spaces and conference rooms where people actively used connected devices.

Modern infrastructure systems operate very differently. Many of today’s most important connected devices are located outdoors, around the perimeter of properties, inside garages or utility areas, through dense walls or across larger physical environments.