Verizon Wins the Quarter and Loses the Tape
A decade-best subscriber quarter, a $4 billion AI Connect book, and a millimeter-wave rescue — all trading 13% below September's high because SpaceX bought spectrum.
VZ · Earnings Call · 2026-10-26
The best quarter nobody is paying for
Verizon's third quarter is, on the operating line, the best it has produced in years. Postpaid phone net additions rose 820% year-over-year to 406,000 — the strongest third quarter in five — while net account additions of 137,000 were the best in over a decade. Postpaid phone churn fell four basis points to 94 basis points. Acquisition cost fell 41%, retention cost 30%. Adjusted EBITDA margin hit 40.2%, up 240 basis points, which CFO Tony Skiadas called “our highest reported” — Tony Skiadas, Chief Financial Officer (CFO) · 2026-10-26. Free cash flow of $7.8 billion was the best quarter since 2020, and full-year FCF growth guidance was raised to approach 12% from 9%-10%. And the stock is down 2.6% over 90 trading days, 13.2% below its September 15 peak. That gap is the whole story: an operating turnaround being repriced by an exogenous competitive shock.Free lines have become a slur
The most revealing shift is rhetorical. In the second quarter of 2025, Verizon's keyword profile was dominated by the free phone guarantee, price locks, and the Value Guarantee — subsidy as strategy. This quarter, the CEO uses the same concept as a term of abuse: “Net adds are easy if your cost of acquisition goes way up and you give away a free line. I don't consider free lines to even be net adds. I think they're churn waiting to happen.” — Dan Schulman, Chief Executive Officer (CEO) · 2026-10-26 The keyword free line now appears as a negative, and the language has migrated from "shake the tree" to "revenue-positive." This is not new talk. It was set up in January's call with an unusually blunt framing:What is new is that the promised arithmetic has arrived. Dan Schulman warned then that refusing to raise prices without adding value would cost roughly 180 basis points of revenue. This quarter, mobility and broadband service revenue grew 2.9%, up from 2.8% in Q2 and 1.6% in Q1, with management guiding Q4 to approximately 4% and wireless service revenue back to growth. The headwind is rolling off exactly as advertised — a rare case of a guidance bridge that held. Total revenue still only rose 1.5% to $34.3 billion, dragged by equipment revenue down 9% as upgrade volumes fell more than 20%; that is the cost of the model, not a failure of it.The #1 rule of getting out of a hole is stop digging. And we're just not going to do that again.
AI Connect: renting the boom instead of buying it
The genuinely new theme is that Verizon has become a supplier to the AI buildout rather than a spectator. Roughly $4 billion of signed contracts in six months, including a dark fiber agreement with Amazon Web Services valued at over $2 billion, layering on the Google deal from last quarter. Management claims “approximately $4 billion of signed deals, and we're just getting started on that” — Dan Schulman, Chief Executive Officer (CEO) · 2026-10-26, with hundreds of millions from edge computing sites built into former central offices, plus a Corning agreement for more than 80 million miles of fiber through 2032 to guarantee supply. This is a broad theme, not a proprietary insight. The global AI data centers complex is among the largest 360-day market movers, spanning dozens of semiconductor, power, and networking names — and data center references now pepper the calls of European industrials from NEX.PA to BRAV.ST to SECU-B.ST. Consensus, in other words. But the tape has cooled: over the last 90 days that same AI data center cluster is a net decliner. Verizon is signing long-duration, success-based contracts into a theme whose equity pricing is deflating. That is arguably the better side of the trade — the revenue is contracted, not marked to market, and the capital behind it is modest and success-based, in Schulman's words.Millimeter-wave redemption, and the SpaceX problem
The most technically interesting item is the rehabilitation of an asset the market had written off. Verizon holds 1,700 MHz of millimeter wave spectrum — twice its competitors — long dismissed as short-range curiosities. Management now claims a technological breakthrough delivering 1-gigabit fixed wireless in metro areas, targeting “50 million-65 million premise passed with one gig or more of download speed” — Dan Schulman, Chief Executive Officer (CEO) · 2026-10-26, and explicitly beyond the old 8-9 million FWA-by-2028 target. It also offloads C-band capacity, and it costs far less per passing than fiber. That is the bull case. The bear case is a rocket company. October's news flow — SpaceX acquiring a nationwide 800 MHz low-band portfolio for $8 billion — knocked the entire carrier group. Schulman did not dismiss it:He then war-gamed the counter: 65,000-80,000 macro sites, $300-$500 billion, four to six years, and a sixth competitor rather than a fourth. The terrestrial network argument is not new — the July call already walked through satellite beam physics and insisted “there is no backdoor to our MVNOs in any structure” — Daniel Schulman, Chief Executive Officer · 2026-07-24. What changed is tone: from "no discernible impact" to a full defensive briefing. Keywords like SpaceX and "satellite" have migrated from curiosity to core narrative in two quarters.I also operate under the assumption that everything SpaceX says they will do, they will do, and that they will execute flawlessly.