Cisco just released new research that reveals AI traffic for enterprise connectivity is set to triple over the next three years, with 50% of that demand concentrated in Wi-Fi. We are witnessing a dangerous disconnect between AI ambition and physical reality.

We are moving from a world where humans use the internet to ‘look things up,’ to one where machines use the network to ‘get things done.’ And when AI becomes a physical participant in our factories and logistics hubs, the network is no longer just a data pipe. It is part of the enterprise’s central nervous system. If that system is sluggish, AI does not just slow down; it fails.

Concretely, in a modern factory, an AI-coordinated robotic system designed to maximize production could ground an entire assembly line to a halt. A few milliseconds of latency could cause the robotic arms to lose synchronization, turning a high-tech asset into a liability.

Spectrum policy is effectively industrial policy. With 73% of businesses expecting their network to hit capacity limits within 24 months, spectrum allocation has shifted from a technical preference to a core pillar of European industrial competitiveness.

If we want businesses to make the most of AI, we should ensure that they have the necessary spectrum capacity for Wi-Fi to support the high-density, low-latency wireless networks these autonomous systems require. Without it, infrastructure will become the primary constraint on innovation.

Read the full research for more insights.