SpaceX Is Not Buying Verizon. The Repricing That Made It Thinkable Already Happened.

By Brian Newman, Contributing Editor

Verizon CEO Dan Schulman told a Detroit audience on September 30 that some form of artificial general intelligence is six to 18 months away, that quantum computing follows two years after that, and that a century of technical progress is about to compress into a decade. The remark is his, attributed, and strongly worded. It is also a planning assumption rather than a settled technical claim, and his own quantum timeline sits outside what most researchers consider credible.

The same week, the trade press revived a different story, which is whether SpaceX buys Verizon outright. Elon Musk has not ruled it out. Semafor reported in July that current and former Verizon executives were privately worried about exactly that. Verizon has denied it twice, in categorical terms.

Treating this as a will-it-or-won't-it story misses what actually changed. The probability of a control deal inside 24 months is low, likely in the low teens. The fact that serious analysts can run the math and have it come out coherent at all is the real news.

Four Numbers That Make the Rumor Plausible

First, spectrum. SpaceX's EchoStar purchase delivered roughly 65 MHz of terrestrial-capable spectrum for about $19.6 billion. That is a real foothold and still a fraction of what Verizon holds across low-band, AWS/PCS, C-band, and mmWave. Upper C-band spectrum comes up for auction by July 2027, and Verizon is widely expected to be the lead bidder. Buying the incumbent skips the auction and the multi-year build entirely.

Second, the network itself. SpaceX's president, Gwynne Shotwell, has described a small-cell-at-the-dish architecture for terrestrial build-out. AT&T chief executive John Stankey called that approach not viable, arguing it would cost as much as a macro network while raising its own permitting problems. Verizon already owns the expensive part: sites, backhaul, an optimized radio network, and the billing and operations systems that turn spectrum into a working service.

Third, fiber, which is the actual AI trade. Verizon signed a $1 billion dark-fiber agreement with Google in July to connect hyperscaler data centers, and a Corning supply deal reported above $2 billion to keep that build unconstrained through 2032. In a world Schulman describes as AGI-dense, the valuable carrier is the one that can interconnect training clusters and host inference at the metro edge. That is the business he has been building in public for months.

Fourth, the currency to pay for it. Verizon's enterprise value sits near $382 billion, with roughly $194 billion in debt. A control premium puts the check above $430 billion. Two years ago that number would have ended the conversation. SpaceX's June IPO priced the company near $1.75 trillion, with post-IPO marks ranging higher, and Musk paid for EchoStar largely in stock. A number that size no longer rules out a deal of that size. It only makes the deal expensive, not unthinkable.

Verizon is not for sale. Its spectrum, its fiber, and its edge capacity are already being priced as something other than connectivity, and the deal does not have to happen for that repricing to be real.

The Probability Is Low. The Repricing Is Already Real.

A control transaction is unlikely in the next two years. The antitrust review alone could run 18 to 36 months, Verizon's unionized, state-regulated operating model does not fold into SpaceX cleanly, and the categorical denial, while not dispositive on its own, lines up with a CEO who just extended his contract through 2028.

The more probable path, judged at roughly 25 percent against the brief's own scenario weighting, is a commercial arrangement short of control. That includes fiber capacity leased to Starlink ground stations, edge colocation at Verizon facilities for inference workloads, or a roaming arrangement in markets where 65 MHz cannot carry the load. That arrangement would let Verizon monetize the asset Schulman is proudest of without a merger review, and it is functionally the same repricing playing out through a lease instead of a purchase.

The Big Three's own October 1 satellite joint venture, pooling spectrum for direct-to-device coverage under interim chief executive Paul Roth, reads the same way. Three carriers that compete for every subscriber just agreed to cooperate rather than let a single satellite operator define the terms of coverage. That is a defensive response to the identical repricing. When connectivity infrastructure becomes AI infrastructure, the owners of that infrastructure start acting like it is worth defending together.

If the control premium on a Tier 1 carrier's hard infrastructure now pencils out against a trillion-dollar company's own stock, how many other telecom assets are quietly being modeled the same way inside investment banks that will never publish the memo? And when the chief executive running the target company is also the one publicly setting the AGI clock, is he selling a company, or is he the one pricing it for everyone else?

Strategic Signal

This is the fourth edition running that has landed on the same underlying argument at a different layer of the stack. AT&T's AI Gateway located the advantage in routing architecture, not the model. The NVIDIA radio unit story showed RAN vendors making capital-allocation bets on silicon economics. The agentic network operations piece showed the same discipline applied to operational trust, with the approval gate as the real product.

This week the pattern shows up one layer further out. It is at the level of who owns the asset outright. The money is moving toward whoever controls the physical and architectural layers next to the model, and that now includes the control premium the market is willing to price on the infrastructure itself.

Watch for a fiber or edge-colocation arrangement between Verizon and a hyperscaler or satellite operator that falls short of a merger. That is the more probable near-term outcome and it would confirm the repricing without requiring a deal.

Watch the upper C-band auction timeline into 2027 as the next point where this math gets retested with real bids rather than analyst modeling, and watch whether other Tier 1 carriers' infrastructure starts getting valued the same way in trade coverage over the next few quarters, independent of whether any specific deal ever closes.

Brian C. Newman is a telecom and AI strategy consultant, course creator, and former Verizon technology leader with more than 30 years of experience across wireless networks, 5G, network operations, infrastructure modernization, and emerging technologies. He helps organizations understand how AI, connectivity, edge computing, and digital infrastructure are reshaping business operations, real estate, public safety, and customer experience.