AT&T Q3: copper retirement creates a cost-timing test

A smaller copper business still needs a managed exit

AT&T’s October 21 third-quarter results put an execution question beside its connectivity growth: how quickly can it remove the costs of a shrinking copper network? The company expects Legacy EBITDA to become negative after 2027 until it has substantially eliminated the network’s direct costs. That forecast makes the transition period relevant even while the remaining business is profitable.

For an investor, a planned revenue decline can be consistent with strategy and still create an earnings headwind. The useful question is whether the expense base can follow customers out of the old system. A revenue chart alone cannot answer that. The analysis below separates reported results, regulatory milestones and the conditions needed for savings to materialize.

The quarterly arithmetic shows the pressure

The financial schedules report Legacy revenue of $1.519 billion, down from $2.013 billion a year earlier. Direct operating expenses were $1.021 billion versus $1.334 billion. Operating income fell to $498 million from $679 million. The respective revenue and expense declines were 24.5% and 23.5%.

Subtracting those disclosed amounts gives a $494 million revenue reduction against $313 million of lower expenses, leaving $181 million less operating income. This is a comparison of reported quarterly balances, not a measure of cash spent on individual shutdown projects. The segment includes copper-based domestic consumer and business services and their direct operating costs; its profit should not be treated as a complete allocation of every corporate obligation.

A practical tracking sheet would keep revenue lost and expenses removed in separate columns. Combining them prematurely into a single savings estimate would hide the timing mismatch the investor is trying to test.

Stopping new sales is only one milestone

The earnings presentation says more than 90% of wire centers have approval to stop selling Legacy services. A separate measure shows 35%+ approved to discontinue providing those services, effective by late 2026. Those are different permissions, and neither percentage should be restated as the share of the network already physically retired.

Think of the distinction as a sequence of questions. Can the business stop adding obligations? Can existing service end? Have the remaining customers moved? Has the associated infrastructure been decommissioned? Has the expense actually left the accounts? An approval statistic answers only its specified question. It does not supply the answers to all later steps.

That reading also prevents a misleading comparison between a geographic approval percentage and a dollar cost reduction. A wire-center count does not establish how much spending each center represents.

Growth provides an offset, with a separate measurement

The presentation reports Advanced Connectivity EBITDA growth of 7.3% year over year and service revenue growth of 6.1%. Those results provide constructive evidence alongside the Legacy decline. They do not demonstrate that copper retirement savings have already been realized. EBITDA is a non-GAAP measure, and segment growth should be evaluated with its definition intact.

The earnings release explains that removing fixed costs across a geography depends on decommissioning infrastructure after customers have upgraded. Approvals could push decommissioning beyond 2029. Accordingly, a model that assumes costs vanish immediately when revenue disappears would run ahead of the company’s stated constraints.

What would change the assessment

The next useful evidence is a connection between operational milestones and reported expenses. Faster completed migrations followed by sustained cost reductions would strengthen the case that the exit is progressing. Revenue erosion without comparable expense removal would increase the transitional burden, even if the broader connectivity franchise keeps growing.

Keep three dated observations in the research record: the exact permission granted, management’s description of completed operational work, and the subsequent expense result. Preserve any qualifications about timing or scope. This approach leaves room for progress without converting an authorization into a realized financial benefit. It also makes the thesis falsifiable when the next quarterly disclosures arrive. If management changes the segment definition, preserve the old figures and use any recast comparisons before extending the trend.