SpaceX

Q2 2026 Results at a Glance

Space Exploration Technologies Corp. (SpaceX) has evolved far beyond a standard aerospace contractor into an end-to-end infrastructure giant. On the orbital transit side, the company flies commercial satellites, runs cargo missions and transports National Aeronautics and Space Administration (NASA) crew to the International Space Station (ISS) using its flight-proven Falcon 9 and Falcon Heavy vehicles. Simultaneously, it is testing Starship, a massive super-heavy lift system designed to achieve full reusability, slash launch costs per kilogram and establish permanent logistics routes to the Moon and Mars.

Beyond physical launch services, SpaceX dominates low-Earth orbit (LEO) telecommunications through Starlink, its constellation delivering high-speed internet to rural communities, aviation lines and maritime vessels worldwide. Its specialized defense division, Starshield, provides national security agencies with secure communications, tactical data relay and Earth observation capabilities.

The operational network extends directly into high-density compute and digital media. Through its integration with the xAI artificial intelligence (AI) ecosystem, SpaceX ties its orbital assets to the Grok language model and the X social media platform, creating a direct loop between real-time data and consumer distribution. To fuel these compute requirements, the company operates the Colossus and Colossus 2 supercomputer data centers in Memphis, Tennessee. Bottom line: SpaceX owns the freight truck, builds the highway, supplies the cargo and runs the AI brain directing traffic.

Falcon 9 / Falcon Heavy

Falcon remains the industrial backbone of SpaceX and the launch system that turned rocket reuse from an experiment into routine infrastructure. Falcon 9 carries commercial satellites, Starlink spacecraft, NASA missions and U.S. national-security payloads while routinely flying previously used first stages. By mid-July 2026, SpaceX reported more than 670 completed Falcon missions, more than 630 booster landings and approximately 600 reflights across the Falcon fleet. Falcon 9 can deliver approximately 22.8 metric tons to low Earth orbit in an expendable configuration. Falcon Heavy, which combines three Falcon-derived first-stage cores, can place nearly 64 metric tons into low Earth orbit and remains one of the world's highest-capacity operational launch vehicles. SpaceX completed 165 Falcon launches in 2025, reinforcing an operating model built around high flight cadence, booster recovery and repeated reuse rather than treating each rocket as a disposable vehicle.

Starlink

Starlink has evolved from a satellite-broadband experiment into SpaceX's largest recurring-revenue business and the world's largest low Earth orbit communications constellation. SpaceX reported more than 12 million active customers across more than 160 countries and territories by June 2026, serving residential, business, maritime, aviation and government markets. The network is built around thousands of maneuverable satellites linked by optical inter-satellite communications, allowing traffic to move through the constellation rather than relying exclusively on nearby ground infrastructure. Starlink Mobile extends the architecture directly to ordinary cellular devices through partnerships with mobile-network operators, creating a path toward satellite coverage in areas where terrestrial towers are unavailable. The next major capacity step is Starlink V3. SpaceX says each V3 spacecraft is designed for approximately 1 Tbps of downlink capacity and 160 Gbps of uplink capacity, roughly 10× and 22× the respective capacity of V2 satellites, with 2,048 beams in each direction. Starship Flight 13 on July 24, 2026 deployed 20 Starlink V3 test spacecraft on a suborbital trajectory, marking the first in-flight deployment demonstration of the new generation. Flight 13 launched July 24, 2026, marking the second flight of the Version 3 vehicle and the program's first deployment test involving Starlink V3 spacecraft. SpaceX also operates Starshield, its government-focused satellite architecture supporting national-security missions.

Dragon

Dragon gives SpaceX something no other U.S. commercial provider has yet matched operationally: a NASA-certified spacecraft routinely transporting astronauts to and from the International Space Station. Crew Dragon was certified by NASA in 2020 and, through Crew-12, has flown 13 crewed missions to the station under NASA's Commercial Crew Program when the Demo-2 test flight is included. Cargo Dragon separately supports NASA's Commercial Resupply Services program, carrying research, equipment and supplies to the station and returning cargo to Earth. Together, Dragon and Falcon 9 restored an independent U.S. orbital human-spaceflight capability after the Space Shuttle's retirement and established SpaceX as a critical transportation provider for NASA's International Space Station operations.

Starship

Starship is the central engineering bet behind SpaceX's next phase. The fully reusable Super Heavy and Starship architecture is being developed to move far more payload per mission than Falcon while eventually recovering and rapidly reflighting both stages. Flight 13 launched July 24, 2026, marking the second flight of the Version 3 vehicle and the program's first deployment of operational Starlink V3 satellites. The mission successfully deployed 20 spacecraft and the Starship upper stage survived its Indian Ocean splashdown intact, although the Super Heavy booster experienced an engine-relight failure during its planned descent and was lost. Starship remains a developmental system rather than an operational replacement for Falcon. Its importance, however, extends well beyond launch. SpaceX intends to use Starship to accelerate Starlink V3 deployment, support NASA's Artemis Human Landing System, move large payloads and infrastructure into orbit and ultimately transport cargo and crews to Mars. The investment thesis is straightforward but execution-heavy: if SpaceX can make both stages rapidly reusable, Starship could fundamentally change the economics and physical scale of what can be placed in orbit. Until that reuse cadence is demonstrated, those cost reductions remain a target rather than an established operating result.

xAI / Grok / X

SpaceX acquired xAI in February 2026, transforming what had primarily been a space-and-connectivity company into a combined aerospace, telecommunications and artificial-intelligence business. The AI operation includes xAI and its Grok family of large language models, most recently Grok 4.5, released in July 2026 and built on a 1.5-trillion-parameter V9 foundation model, trained alongside Cursor, as well as the Colossus computing infrastructure in Memphis and X, formerly Twitter. The acquisition also created a new revenue stream from selling scarce AI computing capacity. Anthropic agreed to pay approximately $1.25 billion per month through May 2029 for access to SpaceX compute capacity, subject to contractual termination provisions, while Google separately agreed to pay approximately $920 million per month from October 2026 through June 2029 for access to approximately 110,000 NVIDIA GPUs and related infrastructure. Those contracts, together with $14.1 billion of additional Cloud Services Agreements signed during Q2 2026, made compute leasing SpaceX's fastest-growing revenue line: AI revenue rose 247% year over year to $2.56 billion in the second quarter, with nameplate compute expanding to 1.4 GW from 0.4 GW a year earlier as Colossus II builds out. SpaceX completed its $60 billion acquisition of the AI coding company Cursor on August 14, 2026. Strategically, the combination links launch, satellite communications, terrestrial computing infrastructure, artificial-intelligence models and a major consumer distribution platform under one corporate structure. Financially, it also exposes SpaceX shareholders to the extraordinary capital requirements and competitive risks of frontier AI.

Terafab

Terafab is a cross-company semiconductor manufacturing project involving SpaceX, Tesla and xAI, all led by Elon Musk. Intel is participating as a semiconductor technology partner, with the manufacturing roadmap centered on Intel's 14A process technology. Public records reviewed by Defense Briefing do not establish a definitive three-company equity allocation, intellectual-property split or revenue-sharing structure, so Terafab should not yet be treated as a conventionally structured three-way joint venture.

SpaceX has the clearest documented legal connection to Terafab's proposed high-volume manufacturing operation in Grimes County, Texas. Texas incentive records identify TeraFab AI, LLC as a SpaceX special-purpose entity, while SpaceX CFO Bret Johnsen executed the company's development agreement with Grimes County. Under that agreement, SpaceX committed to invest at least $5 billion in the county by 2030 and create at least 1,800 full-time jobs by 2035. The agreement also allows SpaceX to terminate with 30 days' written notice. Separate state and school-district incentive applications contemplate a much larger multiphase buildout potentially reaching approximately $119 billion, although that figure remains a planning estimate rather than committed capital.

The manufacturing strategy extends beyond conventional chip fabrication. Terafab's long-term objective is a vertically integrated workflow spanning logic, memory, advanced packaging and related semiconductor production for artificial-intelligence, robotics and space-computing applications. Tesla is leading development work associated with the Giga Texas pilot fab, while SpaceX is positioned as the principal corporate vehicle for the proposed high-volume Grimes County complex. Intended applications include processors for Tesla vehicles and Optimus robots as well as radiation-tolerant computing for SpaceX's orbital infrastructure.

The project has also begun generating evidence of industrial execution beyond company announcements. ASML said in July 2026 that its 2027-2028 capacity planning incorporates anticipated Terafab equipment demand. Terafab nevertheless remains a development-stage manufacturing program. Its ultimate production capacity, yields, capital requirements and timetable have not been demonstrated at commercial scale.

SpaceX's revenue engine is increasingly diversified, but the businesses are at very different stages of maturity. Starlink provides the largest recurring commercial revenue base through residential broadband, enterprise connectivity, aviation, maritime and government services. Falcon contributes launch revenue from commercial customers, NASA and U.S. national-security missions, while Dragon adds long-duration NASA crew and cargo contracts. The xAI acquisition added advertising, subscriptions, artificial-intelligence products and, increasingly, large-scale compute leasing.

The next phase of growth depends less on simply launching more Falcon 9 missions and more on whether SpaceX can make its enormous infrastructure investments reinforce one another. Starship is expected to carry much larger batches of Starlink V3 satellites, which could sharply increase network capacity. Greater Starlink capacity can support additional broadband, enterprise and direct-to-device customers. AI infrastructure can generate contracted compute revenue while supporting Grok and other internal workloads. Terafab could eventually reduce dependence on outside semiconductor manufacturing. In theory, each business lowers a constraint faced by another.

That vertical integration is also the central financial risk. SpaceX is simultaneously funding Starship development, a rapidly expanding satellite constellation, terrestrial AI data centers, artificial-intelligence model development and semiconductor ambitions. The Anthropic and Google compute agreements provide potentially substantial contracted revenue, but the company must convert that revenue and its IPO capital into durable cash generation rather than allowing new infrastructure programs to consume it. For investors, the question is no longer whether SpaceX can build rockets or operate a satellite network. Both have been demonstrated at scale. The harder question is whether a company attempting to industrialize launch, telecommunications, artificial intelligence and computing infrastructure at the same time can produce returns commensurate with the capital required to build them.

Q2 2026 supplied the first quantified read on that question. AI became the fastest-growing line in the company, up 247% year over year to $2.56 billion, driven almost entirely by compute leasing: $14.1 billion of Cloud Services Agreements signed in the quarter delivered $1.6 billion of incremental infrastructure revenue immediately. Connectivity grew 66% to $4.29 billion, with enterprise and government revenue up 108% and consumer up 44%, while Starlink ARPU stopped declining. Space grew 29% to $962 million on a more favorable customer-launch mix even as total launch count fell. The mix shift is significant for how the company should be valued: SpaceX now derives roughly a third of revenue from AI and more than half from Connectivity, with launch services under 9% of the total. The aerospace company is increasingly an infrastructure company whose launch capability is a captive input rather than the principal product.

The first public quarter answered the growth question and replaced it with a capital question. Revenue of $7.81 billion beat consensus near $6.9 billion by roughly $900 million, the operating loss narrowed to $143 million from $970 million a year earlier, and Adjusted EBITDA nearly tripled to $3.54 billion. Every segment grew. The stock still fell roughly 9% after hours, because capital expenditures came in at $18.37 billion against an average estimate near $13.2 billion, and CFO Bret Johnsen guided to similar levels for the next two quarters. SpaceX spent more than twice its quarterly revenue on capital in a single quarter. That is the number that now governs the equity story.

Segment by segment, the picture is coherent. Connectivity is the engine: $4.29 billion of revenue, $1.66 billion of operating income, 12.0 million Starlink subscribers doubling year over year, and ARPU stabilized at $66 after four quarters of decline. Enterprise and government connectivity revenue rose 108% year over year to $1.81 billion, aided by more than $6 billion of multi-year Starshield awards from two Space Force programs, the American Airlines agreement, new airline activations and direct-to-device partnerships with SoftBank, NTT Docomo and Spark NZ. AI is the swing factor: revenue up 247% to $2.56 billion on $14.1 billion of newly contracted Cloud Services Agreements, positive Adjusted EBITDA of $1.15 billion, and an operating loss cut 49% sequentially, but $15.83 billion of the quarter's capex sat in that one segment. Space is the investment: revenue up 29% to $962 million while the operating loss widened to $542 million as Starship research and development accelerated.

On the webcast, Elon Musk said he considers the Starship heat-shield problem “solved at this point” based on Flight 13 data, suggested daily Starship launches were possible roughly a year out, and described an aspiration of moving well over a million tons to orbit annually. Johnsen said SpaceX is on pace for $100 billion of annualized recurring revenue by year end and defended AI compute capital as functionally closer to cost of goods sold given a payback of under one year. Those two framings deserve different weight. The recurring-revenue pace and the payback claim are testable within a few quarters. The cadence and tonnage figures are ambition, and SpaceX's own Q2 data cuts against near-term optimism: total launches fell to 38 from 46 a year earlier and mass to orbit fell 26% as the fleet transitions.

What to watch next quarter. First, whether AI capex stays near $16 billion and whether contracted cloud revenue converts at the margins management implies, the sub-one-year payback assertion is the single largest unverified claim in the release. Second, whether Starlink ARPU holds at $66 now that it has stopped falling; stabilized ARPU on a doubling subscriber base is what turns Connectivity from a growth story into a cash engine. Third, whether the $60 billion Cursor acquisition closes on schedule in Q3 and how it is funded. Fourth, Starship reuse: Flight 13 deployed 20 V3 satellites and survived reentry intact, but booster recovery remains unfinished. Fifth, the balance sheet. SpaceX ended the quarter with $100 billion of cash and securities against $39.4 billion of debt and finance leases, and generated $3.47 billion of operating cash in the first half against $34.5 billion of investing outflows. The company is funding an industrial buildout out of an IPO and a bond deal, not out of operations. That is a legitimate strategy with a finite clock on it.

SpaceX's moat is not one technology. It is the feedback loop between several of them. Falcon's launch cadence creates flight heritage and spreads fixed infrastructure costs across more missions. Reuse lowers the need to manufacture a new first stage for every launch. Starlink provides an internal customer capable of filling launch manifests at a scale no outside satellite operator can match. Starlink then produces recurring connectivity revenue and gives SpaceX a reason to continue pushing launch cost and capacity lower. Dragon adds NASA human-spaceflight credentials while national-security missions deepen government relationships. Starship, if it reaches rapid reuse, could strengthen every part of that loop by carrying substantially more payload per mission.

The xAI acquisition adds another layer but its moat is less established. SpaceX now controls a frontier AI model, the X distribution platform and gigawatt-scale computing infrastructure, while contracted compute agreements create a path to monetize capacity. That can become strategically valuable, particularly if SpaceX eventually integrates terrestrial AI infrastructure with orbital communications and computing. It also makes the company harder to analyze. The aerospace and Starlink businesses benefit from years of demonstrated operational advantage. The AI business remains capital intensive, loss-making and exposed to a much faster competitive cycle.

The primary operating risk remains Starship. SpaceX needs the vehicle not merely to fly but to become reliably reusable at a cadence capable of supporting large-scale Starlink V3 deployment, Artemis requirements and the economics management ultimately expects from the system. Starlink faces a different problem: subscriber growth remains exceptional, but average revenue per user is falling as the customer mix becomes more international and lower-priced. Connectivity must therefore generate sufficient operating leverage to offset declining revenue per subscriber. AI introduces another set of risks. The segment remains deeply loss-making, requires enormous amounts of capital and power and competes in a market where hardware, models and customer preferences can change rapidly.

Q2 2026 sharpened the capital risk specifically. Capital expenditures of $18.37 billion in one quarter represented more than twice quarterly revenue, with 86% directed at AI infrastructure. Management guided to similar spending for the next two quarters and defended it on the basis of a sub-one-year payback on compute capital, a claim that has not yet been demonstrated across a full cycle and that depends on cloud demand, power availability and GPU pricing all holding. The completed $60 billion Cursor acquisition adds a further large commitment before that payback thesis is proven. Meanwhile the operating businesses generated $3.47 billion of cash in the first half against $34.5 billion of investing outflows, meaning the buildout is being funded from IPO proceeds and debt rather than from operations. Launch metrics also moved the wrong way: 38 launches and 485 metric tons to orbit in Q2, down from 46 launches and 652 tons a year earlier, during the Falcon-to-Starship transition.

Financial risk increased after the IPO rather than disappearing. SpaceX raised an extraordinary amount of equity capital and then issued another $25 billion of senior debt in June. The company can fund infrastructure at a scale most competitors cannot approach, but shareholders are now underwriting launch infrastructure, Starlink satellites, Starship, terrestrial AI data centers and semiconductor ambitions simultaneously. Governance and key-person risk remain material because Elon Musk retains unusual influence over strategy while the combination of aerospace, telecommunications, artificial intelligence and X creates regulatory exposure across multiple jurisdictions and agencies.

Control and Beneficial Ownership

Section 16 Activity

Founded in 2002, SpaceX began with a narrowly defined but unusually difficult objective: reduce the cost of reaching orbit enough to make large-scale human activity beyond Earth economically possible. Falcon 1 proved the company could reach orbit. Falcon 9 and booster recovery changed the economics of its launch business. NASA's Commercial Cargo and Commercial Crew programs gave SpaceX an anchor government customer and helped establish Dragon as an operational transportation system. Starlink then changed the company again, turning launch capability into the infrastructure layer for a global communications business rather than merely a service sold to outside satellite operators.

The February 2026 acquisition of xAI marked another structural break. SpaceX absorbed the Grok artificial-intelligence business, X and large-scale Colossus computing infrastructure, pushing the company beyond aerospace and telecommunications into frontier AI and compute services. SpaceX subsequently entered major compute-leasing agreements and continued to invest heavily in AI infrastructure. In August 2026 it completed the $60 billion acquisition of Cursor, adding a major AI software business and issuing approximately 391 million Class A shares as merger consideration.

The June 12, 2026 initial public offering completed the company's transformation from closely held aerospace disruptor into one of the world's largest publicly traded technology companies. SpaceX priced its IPO at $135 per share. Public ownership brings substantially greater access to capital, but it also exposes the company to quarterly financial scrutiny, public-market valuation pressure and direct investor judgment of projects that can require years of spending before producing meaningful revenue. The company's S-1 filing underscored that tension: SpaceX generated approximately $18.7 billion in revenue during 2025, up roughly 33% from 2024, but reported an accumulated deficit of approximately $41.3 billion as of March 31, 2026. That figure is an accounting measure reflecting cumulative retained losses and other adjustments, not simply a cash tally of money lost since 2002. SpaceX therefore enters its public-company era with an unusual combination of operational dominance, enormous infrastructure requirements and businesses ranging from mature launch services to highly speculative bets on reusable super-heavy launch, artificial intelligence and semiconductor manufacturing.

The first public earnings report on August 4, 2026 gave that structure its first outside audit. Second-quarter revenue reached $7.81 billion, up 92% year over year, with all three segments growing and the consolidated operating loss narrowing to $143 million. The IPO and the subsequent $25 billion bond issuance left SpaceX with roughly $100 billion of cash, cash equivalents and marketable securities and a $47.5 billion backlog at quarter end. The accumulated deficit stood at $41.9 billion. Capital expenditures of $18.37 billion in a single quarter, $15.83 billion of it inside the AI segment, demonstrated how aggressively management intends to deploy that balance sheet.