Verizon raised guidance a third straight quarter without raising capital spending — the AI backhaul bill is being paid by copper retirement

Verizon raised full-year guidance a third straight quarter without moving its capital plan, holding 2026 capital spending at the $16.0 billion to $16.5 billion range set in January. Third-quarter net income of $5.3 billion, adjusted earnings of $1.31 a share and free cash flow of $7.8 billion, up 12.1%, gave management room to lift free cash flow growth guidance to approximately 12% from 9% to 10% and to raise the postpaid phone net addition target by 100,000. Capital returned to shareholders reached $12.8 billion in the first nine months, up 49% year over year, with up to $4.5 billion of buybacks expected for the full year.

Underneath the guidance, the mix is shifting: Business segment EBITDA rose 13.0% to $1,941 million while Consumer mobility and broadband service revenue grew just 3.6%. Consumer segment EBITDA rose 6.5% to $11,965 million on a 45.3% margin, and Business margin widened to 27.6% from 25.1%. Verizon's Consumer and Business other-revenue lines grew 6.0% and 10.0%, which runs against the legacy run-off AT&T is executing — AT&T's legacy segment revenue fell 24.5% year over year in the same quarter. The likely reason is timing: Frontier entered Verizon's accounts on January 20, 2026, so Verizon is buying legacy revenue today and harvesting it later.

AT&T's copper customer base has shrunk to 1% of eligible customers from 5% of residential customers in December 2024. More than 100 wire centers are already shut, and copper service has been disconnected in over 500 communities covering 1.3 million locations across 18 states, with another 1,500 communities slated for decommissioning over the next year. AT&T has Federal Communications Commission approval to retire more than 30% of its wire centers by the end of 2026, and expects about half of its wire centers outside California to stop serving customers by the end of 2027. The prize is $4 billion of annual consolidated cost savings by the end of 2028.

On October 6, 2026, AT&T, Global Infrastructure Partners and CPP Investments agreed to form a wholesale, open-access fiber joint venture that combines Forged Fiber 37 — the business AT&T bought from Lumen for $5.75 billion in cash in February — with Gigapower, its existing wholesale venture with GIP. AT&T keeps 50% and hands the other half to infrastructure and pension capital, with closing expected in the first half of 2027 and proceeds going to AT&T toward a net debt target of 2.5 times during 2029. That is a different funding model from spending out of operating cash flow: the fiber gets built, the asset gets partially monetized, and the carrier keeps a long-lived contracted position.

The demand side is what makes the model work, and it is getting larger. Amazon, whose AWS unit is named in the Corning agreement as the kind of hyperscaler the fiber must connect, lifted capital investment to $173 billion on a trailing twelve-month basis through June 2026 from $83 billion in fiscal 2024. At that pace, the Verizon–Alphabet dark fiber agreement amounts to less than 1% of one year of Amazon's capital spending — a rounding error for the buyer, a contract worth more than a billion dollars to a carrier whose entire 2026 capital budget is $16.0 billion to $16.5 billion.

Over the two years to mid-2026, Corning grew EBITDA 70% and Ciena 176%, against 1% at Verizon and 4% at AT&T. The fiber chain compounded while the carriers barely grew, each figure measured from fiscal 2024 to the most recent trailing twelve months. On revenue the gap is the same: Corning is up 29% and Ciena 50%, against 3% at Verizon and 5% at AT&T. Amazon is up 22%, and Lumen is down 10% with EBITDA nearly halved. The carriers are harvesting old copper; the fiber chain is getting paid for new build.

Two caveats keep this from being a simple supplier story. The carrier numbers are flattered by consolidation — Verizon's trailing revenue includes only part of a year of Frontier — and the supplier numbers depend on a build cycle that is contracted rather than speculative: the Corning agreement runs 2027 through 2032, and AT&T's multi-year selection of Ciena for metro and long-haul optical transport covers the equipment that expands and maintains that network. On spending, AT&T's trailing capital expenditures of $22.4 billion are higher than Verizon's $18.6 billion, and both budgets are flat or shrinking. Success-based spending means carriers build only against signed demand, which caps overbuild risk and caps upside too.

The asymmetry between the two carriers decides how fast this converts into cash. AT&T has approval to retire more than 30% of its wire centers by the end of 2026 and can begin discontinuing service in California on or after June 1, 2027. Verizon, which inherited one of the country's largest remaining copper footprints with the Frontier acquisition, filed on September 8, 2026 to retire copper at all remaining locations in 62 New Jersey wire centers — and late in September the FCC pulled the broader nine-state request out of expedited review after state and local officials objected.

Lumen is the template, and it is a warning. After selling substantially all of its consumer fiber-to-the-home business to AT&T for $5.75 billion in cash on February 2, 2026, it kept its national, regional and metro fiber backbone, its central offices and real estate, its copper-based consumer services, and its enterprise and wholesale fiber customers. Reported earnings fell before the cash arrived: revenue dropped from $13.1 billion in fiscal 2024 to $11.8 billion on a trailing twelve-month basis through June 2026, EBITDA from $4.1 billion to $2.2 billion, and the company posted a $1.0 billion net loss. The balance sheet then improved — roughly $4.8 billion of proceeds and cash on hand retire all super priority debt, cutting annual interest expense by about $300 million and taking total debt below $13 billion.

The near-term catalysts are dated; the structural shifts are not. Watch the sequencing, because the cash flows arrive in a different order from the costs — legacy revenue leaves first, cost savings build through 2028, and contracted AI revenue has not yet surfaced as a disclosed line in reported results.

The synthesis: US carriers are not funding the AI backhaul build with capital expenditure. They are funding it with three things — copper cost removal, the sale of legacy assets into joint ventures and to infrastructure capital, and hyperscaler contracts that make new build self-financing. Verizon's third quarter is the cleanest proof: guidance rose while capital spending stayed flat below every year since 2023, and free cash flow growth guidance went to about 12%. The trade is the fiber and optical suppliers and the contracted buyers of capacity, not the carriers' build programs. Two things would falsify it: capital spending guidance rising without a signed demand contract behind it, and a further slip in Verizon's copper filings. And the suppliers carry the cyclical risk — Amazon's trailing capital investment of $173 billion is the demand base underneath Corning's and Ciena's order books, so a pause in hyperscaler spending hits them before it hits the carriers.