Verizon Repurposes Copper Sites for AI Inference

Verizon is repurposing thousands of central offices previously dedicated to copper telephone networks into small data centers for AI inference. CEO Dan Schulman indicated that the company is leveraging the existing power permits, backup systems, and network connections at these sites to capture emerging demand. An initial trial of this infrastructure sold out its available capacity within 24 hours, prompting Verizon to sign new agreements for additional locations.

The strategic shift is driven by the scarcity of usable electrical capacity in the current energy market. By utilizing infrastructure already built to consume high levels of electricity, Verizon aims to generate noticeable revenue from inference and edge computing services next year. This move represents a pivot from viewing energy solely as a cost center to recognizing it as a monetizable asset within the telecommunications business model.

Energy Consumption Drives Strategic Shift

Mobile operators consumed approximately 310 terawatt-hours of electricity in 2025, accounting for roughly 1% of global usage and incurring energy costs of around $52 billion. Networks account for more than three-quarters of this electricity consumption. Historically, operators focused on reducing radio access network consumption and replacing diesel generators with efficient grid power to lower operational expenses.

The economic landscape is changing as the same infrastructure designed to consume power now generates value. Base station batteries are being integrated into electricity markets, and tower companies are developing power-related businesses. The difficulty in rapidly obtaining new electrical capacity for AI workloads is pushing telecoms to monetize their existing energy footprints rather than merely reducing them.

Battery Management Creates New Revenue

Elisa in Finland is managing over 2,000 telecom sites and more than 150 megawatt-hours of distributed battery capacity through its Gridle platform. This virtual power plant model allows the company to optimize battery usage by balancing network resilience requirements with participation in wholesale and balancing electricity markets. The software determines how much charge must remain reserved for backup while selling excess capacity to energy buyers.

This approach transforms idle backup assets into active revenue generators. Software developed by telecoms to manage internal battery storage is now being sold to the broader energy industry. While telecom operators are not becoming utilities, energy management is emerging as a distinct infrastructure business line. The trend reflects a broader industry adaptation where legacy electrical infrastructure acquires new economic utility in the AI era.