technology

At&t's 2025 report lands with a three-month delay and a 40m-fiber punch

While most telcos were already pitching 2027 roadmaps, AT&T finally dropped its 2025 scorecard on a sleepy spring Monday. The headline: 1.5 million new postpaid phones, the eighth straight year above a million fiber adds, and a quiet promise to light up 40 million locations with glass by the end of 2026.

The fiber sprint turns into a marathon

CEO John Stankey calls it “the best consumer broadband growth in a decade,” but the numbers feel more like gravity-defying inertia. The network now passes 32 million homes and businesses; the target is 60 million by 2030. Do the division: that is an extra 2.8 million new locations every year for the next five years, or roughly one household every 11 seconds if the crews work 24/7.

Converged customers—people who bundle fiber with wireless—are the golden geese. They churn less and spend more, so AT&T is essentially bribing itself to keep digging trenches. The more fiber it lays, the stickier its base becomes. Capital intensity stays sky-high, yet investors yawned; the stock barely moved. Wall Street has seen this movie before.

Satellite text messages are coming from space

Satellite text messages are coming from space

Buried beneath the fiber fanfare is the beta launch of satellite-to-cell service, piggybacking on AST SpaceMobile’s orbiting rigs. No firm date, just “2026,” which in telecom speak means anytime between January and the company Christmas party. First responders on FirstNet get first dibs, then the rest of us can text from the middle of nowhere—assuming the FCC signs off and the satellites stop wobbling.

That is the same playbook T-Mobile is running with Starlink, so the race is officially orbital. Whoever cracks voice first wins bragging rights; for now, both carriers are stuck at 2G-style SMS speeds hurtling 500 km above Earth.

Ai is the new axe for costs

Ai is the new axe for costs

AT&T saved $1 billion last year by shoving more customer service chats into bots and letting algorithms predict which cell site will fry next. The new goal: $4 billion annually by 2028. Translation: fewer human technicians, more dashboards that turn red before anything breaks. If you work in operations, your Slack probably already has a channel called #automation-updates that no one volunteers to moderate.

The company insists the savings will be “AI-driven, not head-count-driven,” but history shows the verbs change, the pink slips don’t. Expect quiet layoff rounds tucked into earnings footnotes.

Wireless growth hits a plateau

1.5 million postpaid adds sounds heroic until you realize the entire industry added roughly the same number. AT&T is merely running in place, swapping customers with Verizon and T-Mobile every quarter. ARPU—the average revenue per user—inched up 1%, barely beating inflation. The real money is still in fixed broadband, where margins sit comfortably above 40%.

Stankey keeps telling analysts that 5G standalone core and network slicing will unlock new revenue. So far, consumers only see a slightly faster TikTok feed and a higher bill.

Bottom line: AT&T’s 2025 report reads like a giant fiber IOU. If the company hits 60 million passings by 2030, it will own the most extensive wireline footprint in the country. Miss by even 10%, and the debt load starts looking radioactive. For now, the only thing actually in orbit is the PowerPoint.