Verizon pivots to AI infrastructure, deploying $1B dark fiber network for Google and converting local offices into edge data centers

Verizon is making a major push into AI infrastructure by leveraging its existing telecom assets. The company announced a partnership exceeding $1 billion to connect Google data centers using Verizon's dark fiber network, while simultaneously converting its own central offices into mini data centers aimed at edge computing workloads.

New 'AI Connect' Strategy#

The telecom unveiled these initiatives under a new business unit called AI Connect, detailed during its second-quarter 2026 earnings call. Verizon leadership signaled this is just the beginning: CEO Dan Schulman indicated the company expects to land multiple additional AI-related agreements totaling "multiple billions of dollars in revenue over the next several years" by year-end.

The Connectivity Challenge#

As AI model deployment drives explosive demand for computing power, the challenge extends beyond raw compute—enterprises need reliable, high-bandwidth connections between geographically distributed data centers. According to Schulman, Verizon sees a major opportunity in optimizing this "data center to data center" connectivity to handle surging computational needs.

Edge Computing and Low-Latency Inference#

Verizon is retrofitting existing central offices—the network hubs that support Internet and mobile service—into what it calls "remote data centers." These facilities would host inference workloads requiring minimal latency, enabling time-sensitive AI applications like autonomous robotaxis, robotics, remote surgery, and self-driving vehicles to operate from the network edge rather than routing all requests to distant hyperscale facilities. By positioning compute closer to users, Verizon aims to reduce delays critical for real-time autonomous systems.