AT&T is turning automation into the operating system for a sweeping corporate transition. The US telecom group is using AI to help run networks, serve customers and write code while it abandons aging copper infrastructure and prepares for a smaller workforce.
The strategy connects three changes that are often discussed separately: software replacing specialized hardware, fiber and wireless replacing landlines, and automated systems taking over work previously performed by people. Together, they show how AI can reshape an established industrial company rather than simply add another digital product.
Automation Is Moving Into the Network Core
AT&T already uses AI for customer service, cell-tower placement and maintenance. Its GeoModeler system can adjust the network during extreme weather, while coding tools and automated diagnostics are intended to resolve service problems faster. The company handles about 15 percent of global internet traffic, so even incremental operational changes occur at unusual scale.
The next phase is more structural. AT&T plans to replace hardware at thousands of central hubs with cloud-based software from DriveNets beginning next year. Software-defined infrastructure can be adjusted remotely, reducing the need for technicians to travel to facilities and manually reconfigure equipment. It also shifts expertise away from maintaining dedicated boxes and toward governing software and automated processes.
Management presents that shift as a route to a more adaptable network. The source does not establish how reliably every automated system will perform or how many roles each deployment will eliminate. It does, however, make clear that the company expects fewer middle managers and junior developers as AI assumes more routine responsibilities.
A Smaller Workforce Is Part of the Plan
AT&T employed nearly 131,000 people in June after shedding more than half of its workforce over the previous decade. It eliminated about 8,000 positions last year and another 2,100 in the first half of 2026. Company leaders say the headcount will not be the same five years from now, although they have not published an official target.
One person familiar with the planning suggested the workforce could fall to 85,000 by 2030. AT&T called that figure inaccurate, so it should not be treated as a company forecast. The verified direction is narrower: additional reductions are planned, and automation is expected to change which skills remain valuable.
The chief technology officer described future roles centered on developing and governing AI agents and supervising automated processes. Physical work will not vanish at the same pace. Fiber technicians will still be needed to install and repair connections in the field. That split illustrates a practical limit of the transition: software can coordinate infrastructure, but much of the network still exists in streets, buildings and equipment rooms.
Retiring Copper Changes the Cost Base
AT&T is also decommissioning copper landlines and DSL, a move that has prompted regulatory and local resistance. By the end of the year, it expects no longer to be required to offer copper service across more than 85 percent of its footprint. The company is concentrating investment on fiber and wireless networks while satellite providers add another competitive pressure.
The retirement program has an energy argument as well as a financial one. AT&T says its energy use fell 13.1 percent between 2020 and 2025. Since 2024, the shift away from copper has saved 660,000 megawatt-hours, which the company compares with the annual electricity consumption of 65,000 homes. Those are corporate figures, and the source does not independently audit the calculation.
The combined program is therefore not simply an AI initiative. It is a redesign of labor, equipment and service delivery around software-controlled infrastructure. Customers may experience the result as faster repairs or different connection options; employees will experience it as a redefinition of which work the network still requires.
