5 Billion Eyes: What the World Cup’s Network Teams Know About Zero-Outage Operations

One team, everywhere at once

The 2026 World Cup runs on a single NOC responsible for infrastructure across 16 host cities and three countries. Every venue has its own topology. Every host country has its own carrier relationships and redundancy model. The environments are heterogeneous, the dependencies are complex, and when something breaks, it has to be diagnosed and fixed fast. There is no halftime for a live broadcast feed.

That’s only possible if you have a complete, live picture of the entire network: every venue’s topology, every carrier relationship, every dependency between systems. Not a diagram drawn at deployment. Not a configuration snapshot from last week. The network as it is right now, queryable, traceable, and current across every domain.

Most enterprise network teams are running the same problem at smaller scale: complex, multi-domain infrastructure where a single team is responsible for everything. The question is whether they have the network context to diagnose any failure, in any part of it, before the business feels it.

What the alert doesn’t tell you

In the 78th minute of a knockout match, a BGP session drops in the broadcast uplink subnet. The NOC dashboard was green five minutes ago. An alert flashes red telling you something is wrong. Monitoring is not the problem. It’s doing exactly what it’s designed to do: tell you something is wrong. What it doesn’t tell you is which systems are in the blast radius, what paths remain viable, or how long before the broadcast notices. The gap is what happens after the alert: the manual investigation that determines why, what it’s affecting, and what to do next. The clock starts and the race to diagnose the problem means all hands on deck.

This is where monitoring ends and the real work begins. The alert starts the clock. The intelligence determines how fast it diagnoses.

An agentic NetOps platform can make that alert actionable by pairing it with a live model of the network: current state across the control plane, data plane, and management plane, continuously refreshed across every domain and comparing it to your intent to ground every action. A network source of truth means it already has the live context it needed to reason the problem to find which systems are affected, which paths are still viable, what to do and in what order.

The layer monitoring and ITSMs don’t cover

The World Cup engineers closed that gap with something most enterprise teams are still building toward: AI-driven diagnosis grounded in live network context. Actual reasoning over current topology, current paths, and current intent, so that when a BGP session drops in the 78th minute, a ticket can trigger the diagnosis agent to start diagnosing the problem before anyone has to ask where to look.

That capability, agents that work from live network context and surface root cause before an engineer is paged, is what the emerging category of Agentic NetOps is built around. It extends the value of the monitoring and observability tools already in place, rather than replacing them.

Why this matters for your network

Your network isn’t running a World Cup. But your business depends on it the same way a broadcaster depends on the uplink staying live. The tolerance for “we’re looking into it” is shrinking. Alert volume grows. Networks get more complex. Cost of an outage increases. The engineers available to diagnose them don’t.

Network teams are building toward this standard because they have no alternative. Agentic NetOps is how that gap closes: not by adding headcount, but by giving every engineer the context to act like the most experienced one in the room.