AI infrastructure is increasingly becoming a distributed systems challenge, with compute spreading from centralized data centers toward the network edge, devices, and other infrastructure layers. That shift is putting new demands on the networks connecting those resources.

In one of two new articles for RCRTech, Sean Kinney examines Qualcomm’s 6G strategy and its effort to connect what the company describes as a “continuum of compute.” The vision extends from devices through RAN and edge infrastructure to centralized data centers, with connectivity, compute, and sensing increasingly converging around AI workloads.

The second article looks at the optical layer supporting that distributed infrastructure. Sean examines why AI is driving operators toward 400G, 800G, and 1.6T optical networking, while also increasing the importance of deterministic performance, power efficiency, telemetry, and automation. Simply adding bandwidth may not be enough as workloads become more dynamic and distributed.

Taken together, the articles point to a broader change in how networks are being designed for the AI era. Connectivity is increasingly tied to where compute is located, how workloads move between sites, and how efficiently infrastructure can respond to changing requirements.

For network operators, that means the AI infrastructure conversation is moving beyond data centers and accelerators. The network itself — from 6G architectures to long-haul and metro optical systems — is becoming an increasingly important part of the compute infrastructure required to support distributed AI.