Networks are becoming the nervous system for real-time AI

The 2025 Ryder Cup at Bethpage Black needed to track 67 AI cameras, 650 WiFi access points, and tens of thousands of moving people across an open golf course—all in real time. Tournament staff needed instant visibility into ticket scans, weather, GPS-tracked carts, concession sales, and spectator queues. The solution: a central operations hub that turned raw data into decisions fast enough to matter.

This is what modern AI demands from networks. It's not just about speed anymore—it's about making decisions at the edge, where things actually happen, rather than sending everything to a distant cloud.

When milliseconds matter

A self-driving car can't wait for data to travel to a cloud server and back. It needs to brake in milliseconds. A factory floor can't afford the latency of centralized processing. A tournament can't wait for insights about crowd flow or weather impact. This shift is quietly reshaping how enterprises think about infrastructure.

For years, the cloud migration was the default move—send everything central, process it there, get insights back. But physical AI changes the equation. When AI moves off screens and onto streets, factory floors, and event grounds, the old model breaks down. "By the time processing happens in the cloud, the machine has already moved," explains Jon Green, HPE's CTO.

The numbers tell the story. Enterprise Research Group found that 84% of companies are rethinking where they deploy applications because of AI growth. IDC forecasts the AI infrastructure market will hit $758 billion by 2029. This isn't theoretical—it's already reshaping how organizations build.

Networks training themselves

But here's where it gets interesting: AI doesn't just need better networks. AI is also making networks smarter. Networks are data-rich systems—millions of configuration states, thousands of environments, endless patterns about what actually improves performance. That's perfect training material for AI.

AIOps (AI-driven IT operations) is already changing how IT teams work. Today, AI surfaces recommendations that administrators approve with a click. Tomorrow, those systems will test and deploy low-risk changes automatically. The repetitive, error-prone work that has always slowed IT teams down—configuring switches, detecting plugged-in connectors, fixing stuck ports—starts getting handled by the system itself.

"AI isn't coming for the network engineer's job, but it will eliminate the tedious stuff that slows them down," Green says. "You'll be able to say, 'Please go configure 130 switches to solve this issue,' and the system will handle it."

This vision of a self-driving network isn't about replacing humans. It's about freeing them from the grinding work that prevents them from thinking strategically. The network becomes less a thing you manage and more a thing that manages itself—while you focus on what matters.

For businesses trying to scale AI applications, this infrastructure shift is the difference between piloting and actually deploying at scale. The networks built today will determine which organizations can move fast enough to compete tomorrow.