SpaceX Is Not Entering Telecom. It Is Leaving It Behind.
By Brian Newman, Contributing Editor
On June 25, SpaceX President Gwynne Shotwell told IPO roadshow investors that the company plans to launch a retail Starlink Mobile service and may build its own terrestrial wireless network in the United States. The disclosure, first reported by the Financial Times, came two weeks after SpaceX completed the largest IPO in financial history, raising roughly $86 billion at a valuation exceeding $2 trillion.
Most coverage framed this as a satellite company entering the wireless market. That framing misses what is actually happening.
SpaceX is not entering telecom. It is assembling a vertically integrated AI-connectivity-compute company at a pace no traditional operator can match. The timeline across the first half of 2026 tells the story.
In January, SpaceX filed with the FCC for authorization to launch up to one million orbital data center satellites, a constellation designed to move AI compute workloads off-planet. In February, SpaceX closed its acquisition of xAI in an all-stock merger valued at $1.25 trillion, folding the Grok AI model, the X social platform, and xAI's research operations into the same corporate entity that builds rockets and operates the world's largest satellite constellation. By June 8, the company unveiled the AI1 satellite: a 70-meter-wingspan orbital compute node delivering 150 kilowatts of peak AI processing power, roughly equivalent to a single Nvidia GB300 server rack. Three days later, SpaceX went public on the Nasdaq.
The financial architecture is equally revealing. Starlink generated $11.4 billion of SpaceX's $18.7 billion in total 2025 revenue, with 10.3 million active subscribers across more than 160 countries. The company then signed compute infrastructure deals that rival the connectivity business in scale. Anthropic is paying $1.25 billion per month through 2029 for access to SpaceX's Colossus data centers. Google signed a $920 million per month agreement for GPU access running through June 2029. SpaceX is already earning more from renting AI compute than most telecom operators earn from enterprise services.
The spectrum position, while smaller than the incumbents, is real. The FCC approved SpaceX's acquisition of 65 megahertz of exclusive-use, nationwide mid-band spectrum from EchoStar in May 2026. Analysts at New Street Research note that 65 megahertz compares unfavorably to the roughly 1,020 megahertz the three major U.S. carriers collectively hold. The gap is significant for a standalone terrestrial network. It matters far less for a hybrid satellite-terrestrial architecture where more than 650 direct-to-cell satellites already provide global coverage, with FCC authorization to expand the constellation to 15,000.
The carriers see the threat. On May 14, AT&T, T-Mobile, and Verizon announced an agreement in principle to form a joint venture pooling spectrum resources for satellite-based direct-to-device connectivity. The move was unprecedented: three competitors that have never collaborated on core network strategy decided to share spectrum and standardize D2D specifications. The stated purpose is ending dead zones. The unstated purpose is presenting a unified front before SpaceX, and now Amazon, bypass the carrier billing layer entirely.
Amazon's roughly $11.6 billion acquisition of Globalstar in April added a second vertically integrated competitor, complete with an Apple partnership for iPhone and Apple Watch D2D services. Between SpaceX and Amazon, two companies with trillion-dollar-plus market capitalizations, independent launch capabilities, and deep AI infrastructure investments now hold the technical capability to route mobile connectivity around the traditional carrier model.
The argument in four parts:
A. SpaceX now controls satellites, launch vehicles, spectrum, terrestrial data centers, and AI models under a single corporate entity. No traditional telecom operator holds all of those layers.
B. The carrier D2D joint venture pools spectrum, which is necessary, but spectrum alone is not a competitive moat when the challenger owns the launch economics, the satellite constellation, and the compute stack.
C. Oppenheimer projects SpaceX could reach 15 million U.S. subscribers by 2030 and compete in what analysts estimate is a $1.6 trillion U.S. communications market. Even if that projection is aggressive, the structural pressure on carrier margins and negotiating leverage is real.
D. The real contest is not coverage. Coverage is the surface-level story. The deeper question is who controls the intelligence layer between connectivity and compute, which is the orchestration point where network resources meet AI workloads.
Traditional operators have spent the last decade debating whether the transformation is "Telco to TechCo." SpaceX skipped that conversation entirely. It built a TechCo, acquired an AICo, and is now entering Telco from a position of structural advantage.
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.
