SpaceX Is Building a Wireless Carrier

Hey there, bargain hunter. On the evening of August 4, SpaceX sat down for its first earnings call as a public company. Revenue beat. EBITDA crushed estimates. And somewhere in the Q&A, Gwynne Shotwell named AT&T, Verizon, and T-Mobile by name, called their combined revenue roughly $600 billion a year, and said she expected SpaceX to win quite a few of their customers. By Wednesday morning, all three carrier stocks were lower. SPCX was trading at $125.33, well below its $135 IPO price. The market is still sorting out what just happened.

Here is the correct frame: this was not a satellite announcement. It was a terrestrial mobile declaration. The distinction matters enormously, and most early coverage missed it.

Scoreboard

SpaceX Q2 2026 results, the first the company has ever released publicly:

• Total revenue: $7.81 billion, up 92% year-over-year, versus a consensus estimate of roughly $6.93 billion
• Net loss: $541 million, narrowed from $1.01 billion a year ago
• Adjusted EBITDA: $3.5 billion, about 75% above the $2.0 billion the Street expected
• Connectivity (Starlink) revenue: $4.29 billion, up 66% year-over-year, versus a $3.83 billion estimate
• Connectivity operating income: $1.66 billion, the only segment generating an operating profit
• AI segment revenue: $2.56 billion, up 247% year-over-year
• Space segment revenue: $962 million, up 29% year-over-year
• Total capex: $18.4 billion for the quarter, with $15.8 billion allocated to AI infrastructure
• Starlink subscribers: 12 million, doubling from a year earlier, with 1.7 million net adds in Q2 alone
• Starlink ARPU: $66 per month, stable versus Q1, down from $85 a year earlier
• Cash and equivalents: $100 billion at quarter end
• Backlog: $47.5 billion
• SPCX after-hours: fell as much as 8% on capex concerns; trading at $125.33 on August 5
• T, VZ, TMUS: each fell in post-market and premarket trading following Shotwell’s wireless comments

The revenue beat was significant. The capex number was the bruise. At $18.4 billion in a single quarter, capital expenditure ran more than double the quarter’s revenue. Analysts had modeled roughly $13 billion. The gap between what was expected and what was disclosed explains most of the after-hours move in SPCX.

The Real Reason the Carriers Dropped

The market had priced SpaceX as a satellite internet company with carrier partnership ambitions. What Shotwell described on the call was something different: a ground-up terrestrial mobile network, cell towers, small cell nodes, and land-based spectrum, layered on top of an existing satellite constellation. That is not a backup service for dead zones. That is a fourth national wireless provider, built from orbit down.

Shotwell was specific. She confirmed the EchoStar spectrum includes terrestrial components and said SpaceX “definitely intend[s] to build out terrestrial” infrastructure to make Starlink “a true mobile service.” She pointed to the roughly $600 billion in combined annual revenue held by AT&T, Verizon, and T-Mobile as the addressable opportunity. That is the revenue pool SpaceX told investors it intends to access.

The carriers responded publicly with measured skepticism. Verizon said satellite technology remains constrained by physics and is better suited for rural markets. AT&T and T-Mobile noted their existing networks offer superior performance for most consumers today. Neither response is wrong. Neither response changes the fact that a company with 10,200 satellites already in orbit, $100 billion in cash, and FCC-cleared spectrum just told the world it is coming for their subscribers.

Deep Dive: The Spectrum Math Nobody Was Modeling

The foundation was built in two moves. SpaceX acquired 65 MHz of wireless spectrum from EchoStar across two transactions totaling $19.6 billion. The first deal, announced in September 2025, covered AWS-4 and H-block licenses for roughly $17 billion, structured as $8.5 billion in cash and $8.5 billion in SpaceX stock. Two months later, the companies amended the agreement to add EchoStar’s unpaired AWS-3 spectrum for another $2.6 billion in SpaceX stock. The FCC cleared the transfer, handing SpaceX exactly the kind of terrestrial spectrum a standalone mobile network requires.

The quantity matters. SpaceX’s current direct-to-cell operation uses around 5 MHz of bandwidth through local carrier partners. The EchoStar transaction delivers roughly 65 MHz. That is a 13-fold increase in spectral capacity, and it sat dormant until Shotwell described it on the earnings call as a “foundational competitive advantage for Starlink Mobile.”

The satellite layer is already substantial. SpaceX ended Q2 with approximately 10,200 operational Starlink satellites in orbit, serving customers across 167 markets. Those satellites provided roughly 800 terabits per second of total downlink capacity. Starlink added 1.7 million net consumer subscribers in Q2 alone, its best quarter for new customer additions. The near-term priority, per Shotwell, is launching Mobile V2 satellites on Starship before integrating the EchoStar spectrum in 2027. Commercial Starlink Mobile service is targeted by end of next year.

Data Section: The Numbers Behind the Ambition

Connectivity segment (Starlink): $4.29 billion in Q2, up 66% year-over-year. Operating income of $1.66 billion. The only profitable segment. Consumer revenue reached $2.49 billion; enterprise and government revenue more than doubled to $1.81 billion, up 108% year-over-year. ARPU held steady at $66 per month versus Q1 but is down 22% from $85 a year earlier, reflecting international expansion and lower-priced plan introductions in new markets.

AI segment: $2.56 billion in revenue, up 247% year-over-year. Adjusted EBITDA turned positive at $1.1 billion. SpaceX ended Q2 with 1.4 gigawatts of nameplate compute capacity, up from 1 GW in Q1 and 400 megawatts a year earlier. CFO Bret Johnsen said on the call that AI compute capex delivers a less-than-one-year payback, calling it capital that is “moving so quickly” it functions more like cost of goods sold than traditional long-duration investment.

Space segment: $962 million, up 29% year-over-year. SpaceX completed 78 launches in the first half of 2026, delivering 1,041 tons of mass to orbit. Adjusted EBITDA loss of $205 million, reflecting elevated Starship R&D spend.

Balance sheet: $100 billion in cash, cash equivalents, and marketable securities at quarter end. Total debt and finance leases reached $36.8 billion, up from $22 billion three months earlier. SpaceX raised approximately $85.7 billion from its June IPO and issued $25 billion in senior notes, using part of the proceeds to retire a $20 billion bridge loan. The company also carries a $47.5 billion backlog.

Guidance: Johnsen said the company is on pace to reach $100 billion in annualized recurring revenue by the end of 2026, the first formal forward guidance in SpaceX’s 24-year history. In the first weeks of Q3, SpaceX had already contracted an additional $6.7 billion in cloud services revenue ramping from October. Management also pulled forward its $1 trillion annual revenue target to 2030.

Is It Cheap?

SPCX priced its IPO at $135 per share on June 12, implying a market cap of roughly $1.77 trillion. By August 5, the stock was trading at $125.33, below the IPO price, after falling as much as 8% in the session following the earnings release. The 52-week range is $104.83 to $225.64, with the consensus analyst price target sitting at $223.

The valuation question is not simple, bargain hunter. SpaceX is running $18.4 billion in quarterly capex against $7.81 billion in quarterly revenue. The company is spending more than it earns in a single quarter. The bull argument is that Johnsen framed AI compute capex as a less-than-one-year payback investment, which, if accurate, means the spending is not destroying value but compressing it into the near term. The bear argument is that $36.8 billion in debt and finance leases growing at this pace creates balance sheet risk if the revenue ramp disappoints.

On a connectivity-only basis, $4.29 billion in quarterly revenue annualizes to roughly $17 billion. With $1.66 billion in quarterly operating income from that segment, the connectivity business alone is generating a roughly 39% operating margin. That is not cheap, but it is not thin either. The question the market is asking is whether the wireless ambition represents a third leg of durable earnings growth or a capital-intensive distraction that stretches an already expensive balance sheet.

For the telecom names, the valuation read-through is more mechanical. AT&T at roughly $22.91 carries a 4.8% dividend yield. Verizon and AT&T are already hedged through AST SpaceMobile for satellite-to-cell coverage, running through a third party. T-Mobile carries a commercial revenue relationship with SpaceX itself through T-Satellite. Each of the three has historically underinvested in rural coverage, which is precisely where SpaceX can build density first without needing to out-compete 5G urban infrastructure day one.

Who Bleeds, Who Watches

AT&T and Verizon face the cleaner competitive read-through. Both have teamed with AST SpaceMobile for satellite-to-cell service, meaning their satellite hedge runs through a third party whose scale and timeline do not match SpaceX’s vertical integration. SpaceX manufactures its own satellites, launches them on its own rockets, and controls one of the largest low-Earth orbit constellations in existence. Dish Network spent billions on spectrum it never turned into a competitive carrier. SpaceX is not Dish: the rockets are real, the satellites are already up, and the cash is in the bank.

T-Mobile occupies the most complicated position. Its existing T-Satellite service uses SpaceX’s Starlink as a background provider for dead-zone coverage, with T-Mobile owning the customer relationship and SpaceX operating the satellite layer. A direct-to-consumer Starlink Mobile plan inverts that dynamic: SpaceX signs up subscribers, bills them directly, and controls its own brand. T-Mobile CEO Mike Sievert noted on the company’s Q2 2026 earnings call that satellite-to-cell service accounts for about 0.0003% of T-Mobile’s network traffic during peak summer months, a framing designed to minimize the threat. The response from Shotwell is essentially the Mobile V2 satellite roadmap targeting commercial launch by end of 2027.

One adjacent name deserves attention. Bloomberg reported last month that SpaceX engaged Charter Communications in discussions about routing a portion of its mobile traffic over Charter’s terrestrial network. If that arrangement advances, SpaceX gets ground-level density without building towers from scratch, and Charter gets a revenue stream that partially offsets its own broadband competition with Starlink. Cable names warrant monitoring if this develops into a formal agreement.

International carrier agreements are already live. SpaceX launched Starlink Mobile partnerships with SoftBank, NTT DoCoMo, and Spark New Zealand during Q2, allowing Starlink to operate as a silent partner to existing mobile operators in coverage gaps. The EchoStar spectrum approval extends that model into the United States at a scale that permits standalone service without a carrier intermediary.

Bull / Base / Bear

Bull Case for SPCX

SpaceX closes a Charter network-sharing deal, reducing its terrestrial buildout cost and compressing the timeline materially. Mobile V2 satellites launch on Starship on schedule. Commercial Starlink Mobile service reaches customers by end of 2027 as guided. Connectivity operating income scales from $1.66 billion per quarter toward materially higher levels as 65 MHz of EchoStar spectrum comes online. AI capex generates the sub-one-year payback Johnsen described, and the segment turns to a net income contributor by 2028. SPCX recovers toward the consensus analyst target of $223.

Base Case

SpaceX builds a hybrid satellite-terrestrial service that captures rural and underserved markets, the same populations the Big Three have historically underinvested in. Carriers defend urban density with existing 5G infrastructure. Subscriber growth at SpaceX continues at a moderating pace as ARPU stabilizes following the June 2026 price increases. Telecom names trade at a mild disruption discount but do not break down. SPCX stabilizes in the $115 to $140 range as the market waits for Q3 data, particularly the ARPU recovery from the June pricing action.

Bear Case

Terrestrial buildout proves slower and more expensive than guided. FCC imposes buildout obligations tied to the EchoStar spectrum licenses that strain the timeline. Charter discussions collapse and SpaceX must acquire towers, adding cost to an already elevated capex run rate. Free cash flow turns deeply negative for multiple consecutive quarters as debt grows from $36.8 billion. Carriers respond with aggressive rural pricing before Starlink Mobile reaches scale. Analysts at Phillip Capital have already flagged that cash and equivalents could remain flat over four years while debt swells by nearly 150%, resulting in a net debt position of $3.3 billion by 2030. The lock-up expiration on August 6, covering roughly 911.5 million insider shares, adds near-term selling pressure. SPCX tests the 52-week low of $104.83.

Action Plan

Three positioning frameworks, depending on time horizon and risk tolerance.

SPCX: scale-in framework below IPO price. The stock is trading at $125.33, below the $135 IPO price. The $135 level now functions as resistance on any recovery attempt. The lock-up expiration on August 6 covering 911.5 million shares is the most immediate variable: heavy downside volume with price stabilizing is a different signal than low volume with continued drift lower. For a long-horizon buyer who believes the wireless and AI ramps are real, a starter position here, with additional allocation reserved for a test of the $105 to $110 range, is a disciplined approach. Do not size aggressively into a lock-up event. This is a 12-to-24-month deployment horizon, not a 30-day trade.

Telecom pair positioning. T-Mobile’s commercial revenue relationship with SpaceX provides partial insulation that AT&T and Verizon do not have. A relative value framework, long TMUS against short T or VZ, captures the competitive divergence without taking a directional view on the sector. Watch whether the premarket weakness (Verizon and AT&T down more than 2%, T-Mobile down 1.3% as of Wednesday morning) extends intraday on elevated volume, which would suggest institutional reweighting rather than a reflexive after-hours reaction.

Charter and cable adjacency. If the Bloomberg report on SpaceX-Charter discussions is accurate, CHTR becomes an asymmetric situation. A network-sharing deal would be a material positive for Charter while also providing SpaceX a faster path to terrestrial coverage. No formal agreement exists yet. Monitor for confirmation before sizing in.

Cheap Investor Scorecard

Track these 10 items over the next four quarters:

1. Starlink ARPU recovery: Does $66/month stabilize and recover toward $75+ in Q3 and Q4 following the June 2026 price increases? The pricing action’s full effect flows through Q3.
2. Lock-up impact: Does the August 6 expiration of 911.5 million insider shares produce sustained selling pressure, or does the stock absorb the supply?
3. Connectivity subscriber growth: Does the 1.7 million net-add quarterly pace hold? Any deceleration below 1 million is a yellow flag for the wireless ambition’s subscriber ramp.
4. Mobile V2 satellite launch cadence: Are Starship-delivered Mobile V2 satellites reaching orbit on schedule ahead of the 2027 EchoStar spectrum integration?
5. AI capex payback evidence: Does Johnsen’s claim of a sub-one-year AI compute return begin showing up in AI segment revenue acceleration and margin improvement?
6. Charter deal confirmation: Does the Bloomberg-reported SpaceX-Charter network-sharing discussion result in a formal agreement? A confirmed deal is a material positive for both stocks.
7. Carrier defensive response: Do AT&T, Verizon, and T-Mobile accelerate rural infrastructure investment or cut pricing to pre-empt Starlink Mobile?
8. Debt trajectory: Does the $36.8 billion in debt and finance leases continue growing, or does the $100 billion cash position fund capex without further borrowing?
9. Annualized revenue run rate: Does SpaceX reach the $100 billion annualized recurring revenue target by December 2026 as Johnsen guided?
10. SPCX vs. $135 IPO price: Does the stock reclaim and hold the $135 level before the next earnings call? A sustained trade below $135 signals the market is still in “prove it” mode.

Bottom Line

What happened on the evening of August 4 was not a product announcement. It was a capital allocation declaration backed by $19.6 billion of FCC-cleared spectrum, $100 billion in cash, 12 million subscribers, and 10,200 satellites already in orbit. SpaceX told investors, analysts, and the carriers themselves that it intends to build a fourth national wireless provider from orbit down.

The carriers have roughly two years before Starlink Mobile’s next-generation service is positioned to scale in earnest. Whether they use that window to defend with pricing, spectrum acquisition, or deeper satellite partnerships will determine how much of that $600 billion revenue pool SpaceX can realistically access in the near term.

If the Charter deal closes and Mobile V2 launches on schedule: SPCX re-rates and the telecom discount widens. If the buildout slips and the lock-up creates sustained selling: the stock tests $105 and the carriers get a longer runway to respond. Preparation, not prediction, is the only edge that holds. Watch the scorecard items above, not the rhetoric.

The Cheap Investor