SpaceX Files IPO Prospectus with SEC, Unveiling Vision for a Multi-planetary and AI-Powered future
In a landmark filing last night, the U.S. Securities and Exchange Commission (SEC) published the full IPO prospectus of Space Exploration Technologies Corp. (SpaceX). The sprawling, 300-plus-page document serves not only as a financial disclosure but as a bold manifesto for the future of space and Artificial Intelligence. Below is the full translated summary of the prospectus.
“You want to wake up in the morning and think the future is going to be great—and that’s what being a spacefaring civilization is all about. It’s about believing in the future and thinking that the future will be better than the past. And I can’t think of anything more exciting than going out there and being among the stars.” — Elon Musk
Our Mission
Our mission is to build the systems and technologies necessary to make life multiplanetary, underStand the true nature of the universe, and extend the light of consciousness to the stars. To achieve this, we have assembled the most ambitious, vertically integrated innovation engine on Earth—and beyond—with unparalleled capabilities: rapidly manufACTuring and launching space-based communication systems that connect the world; harnessing solar energy to power truth-seeking artificial intelligence that drives scientific discovery; and ultimately, establishing a base on the moon and cities on other planets.
Overview
Founded in 2002, SpaceX is the only company operating across three critical sectors—Space, Connectivity, and AI—building integrated hardware and software infrastructure for the future. We are builders at our core. We design, manufacture, launch, and operate products and services based on cutting-edge Technology, including the world’s most advanced rockets and spacecraft. We safely and reliably transport Astronauts, satellites, and other payloads on missions that benefit Earth. Since 2023, our Falcon rockets have launched over 80% of the world’s total orbital mass annually, with a mission success rate exceeding 99%. We also operate a low-Earth orbit (LEO), high-speed, low-latency global broadband data and communications network comprising APProximately 9,600 Starlink broadband and direct-to-mobile satellites, serving millions of consumer, enterprise, and government customers across 164 countries, territories, and other markets as of March 31, 2026. Through our dedicated direct-to-mobile satellite constellation, we provide connectivity services in roughly 50 countries, complementing terrestrial networks and significantly reducing mobile dead zones.
AI has immense potential to improve space exploration and life on Earth, accelerating our mission. xAI, founded in 2023 and acquired by SpaceX in early 2026, now forms an indispensable pillar of our vertically integrated company. We are rapidly building AI Compute infrastructure with industry-leading speed and cost efficiency—starting on Earth, with the goal of extending into space. Our infrastructure supports the training and inference of Grok, which has become one of the world’s most advanced frontier models. Grok is designed as a truth-seeking AI model, grounded in founder Elon Musk’s mission to understand the universe. We believe achieving this mission requires an AI that takes a truth-seeking approach. We DeFine "truth-seeking" as the active, relentless pursuit of objective truths about reality, based on evidence, logic, empirical data, and first-principles thinking. Our goal is to understand and explain how the universe works as accurately as possible. Within two years of its initial model release, Grok reached frontier-level performance in scientific reasoning (measured by GPQA Diamond score, a benchmark of expert-written, validated standardized questions) on a Timeline faster than reported by other leading model providers. Grok also benefits from deep integration with X, our real-time Information, entertainment, and free speech platform, which serves as the core distribution and data engine for our AI ecosystem, further strengthening Grok’s truth-seeking objective.
We believe space represents the largest economic frontier in human history. Space-based connectivity infrastructure aims to grant everyone on Earth access to Education, healthcare, entertainment, and communication, enabling people to transcend traditional limitations like physical and political borders. We believe space-based AI infrastructure can harness the near-limitless power of the sun, making AI a transformative force for understanding the universe and improving daily life for all humanity. We believe the convergence of these domains will unlock an unprecedented expansion of the global Economy, ushering in an age of abundance. Our innovations and technological advancements are redefining industries on Earth, while we simultaneously aim to create new industries on the Moon, Mars, and beyond. We are truly building the infrastructure of the future.
Space. SpaceX is the only company to have cracked the code of large-scale space access, revolutionizing an industry characterized by decades of stagnation, risk aveRSIon, and economically distorted cost structures. SpaceX disrupted this paradigm by applying first-principles thinking—rejecting industry assumptions and building solutions based on the fundamental lAWS of physics. Our intense, mission-driven, engineering-first culture and focus on extreme vertical integration have enabled us to achieve what many deemed impossible. We pioneered frequent, reliable, and affordable space access with the Falcon rocket family. In 2015, we successfully landed the first Falcon 9 booster back from space, beating the rest of the industry by at least a decade. Spaceflight, which historically cost billions per launch, now costs tens of millions of dollars, radically lowering the cost of accessing space and creating opportunities to establish new ventures in orbit.
Connectivity. Since activating service for customers in 2020, Starlink has rapidly expanded high-speed internet access globally, prioritizing severely underserved rural and rEMOte communities worldwide. Building terrestrial networks in such communities is prohibitively expensive, whereas Starlink requires only a single Starlink kit to provide broadband connectivity anywhere on Earth. As of March 31, 2026, we operated approximately 9,600 Starlink broadband and direct-to-mobile satellites in LEO, the world's most advanced broadband constellation, serving approximately 10.3 million Starlink users with internet connectivity in 164 countries, territories, and other markets. In January 2024, we also began deploying the Starlink Mobile constellation, using separate direct-to-mobile capable Starlink satellites to drastically reduce mobile dead zones globally. As of March 31, 2026, our constellation of approximately 650 V1 Mobile dedicated direct-to-mobile satellites provided satellite-to-mobile data, OTT voice, and messAGIng services to roughly 7.4 million monthly active unique devices across about 30 countries.
AI. We are the first company to deploy a gigawatt-scale AI training cluster. For complex reasoning and agentic workloads, compute is directly correlated with the quality of intelligence and the speed of task completion. In less than two years, we have built a dual advantage in both cost efficiency and speed of large-scale deployment. By owning our compute infrastructure and vertically integrating across the entire AI stack, we can train and iterate frontier models at lower costs and higher speeds, accelerating development cycles. This eliminates external bottlenecks and drives rapid, continuous improvement in Model Performance. We believe our combination of state-of-the-art AI compute infrastructure, a truth-seeking frontier model, and real-time data access on X creates a significant strategic advantage. Our integrated AI platform of Grok and X supported over 1.3 billion active accounts in the twelve months ended March 31, 2026, including approximately 550 million MAU, and generated roughly 350 million daily posts. Among our MAU, approximately 117 million had used Grok's AI features as of March 31, 2026. The deep integration of Grok with X enables freshness, relevance, and contextual awareness, which we believe are competitive differentiators. This direct, real-time access to the information and human discourse on X enhances Grok’s truth-seeking capabilities by grounding its ouTPUts in the latest knowledge and diverse perspectives. Consequently, we believe Grok can deliver the most objective, relevant insights and best serve high-frequency, high-value use cases in consumer and enterprise AI applications.
We have created unique new markets across the space, connectivity, and AI industries by building integrated future hardware and software infrastructure combined with broad capabilities. For instance, SpaceX’s recent acquisition of xAI combines SpaceX’s launch capabilities and global connectivity network with xAI’s AI Development capabilities. Specifically, we believe SpaceX’s reusable rockets, scaled satellite manufacturing, and Operational expertise can enable the cost-effective and rapid deployment of a large-scale AI compute satellite constellation—potentially comprising millions of satellites—for orbital data centers. We believe these AI compute satellites in sun-Synchronous orbit will be able to process energy-intensive AI workloads, such as inference demand, at a scale and efficiency far exceeding terrestrial alternatives, while Starlink provides the low-latency, global connectivity to link these orbital AI systems to people worldwide and provide real-time Intelligence. We expect to begin deploying our orbital AI compute satellites as early as 2028.
Our Financial Performance
Our financial results demonstrate the strength of our operating model and our ability to create and scale multiple new businesses:
For the three months ended March 31, 2026, our consolidated revenue was 1.943 billion, and Adjusted EBITDA was $1.127 billion.
For 2025, our consolidated revenue was 2.389 billion, and Adjusted EBITDA was $6.584 billion.
Our Space and Connectivity segments contributed the vast majority of consolidated revenue in the three months ended March 31, 2026, and the year ended December 31, 2025, demonstrating the power of scale and operating leveRAGe in our vertically integrated business model:
For the three months ended March 31, 2026, our Space segment revenue was (662) million, and segment Adjusted EBITDA was 4.086 billion, operating loss was 653 million. Additionally, our Space segment invested 3.004 billion in R&D for our next-generation Starship launch vehicle program during these respective periods. Starship is designed to achieve a step-change improvement in reusability, payload capacity, and launch cadence, serving as a key enabler of our long-term growth strategy and unlocking entirely new mission categories.
For the three months ended March 31, 2026, our Connectivity segment revenue was 1.188 billion, and segment Adjusted EBITDA was 13.87 billion, operating income of 7.168 billion in 2025, reflecting year-over-year growth of 49.8%, 120.4%, and 86.2%, respectively, benefiting from user growth, increased enterprise adoption, and continued network efficiency improvements.
In our newly acquired AI segment, we plan to prioritize growth and Investment to capture significant opportunities in AI applications and compute infrastructure. For the three months ended March 31, 2026, our AI segment revenue was (2.469) billion, and segment Adjusted EBITDA was 3.201 billion, operating loss was (1.237) billion, reflecting its early development stage and ongoing investments to support AI long-term growth opportunities.
For the three months ended March 31, 2026, capital expenditures were 1.332 billion for Connectivity, and 3.832 billion for Space, 12.727 billion for AI.
Why This Matters Now
For the entirety of human civilization, humanity has lived on a single celestial body: Earth. This current paradigm, where human civilization is confined to a single planet, exposes humanity to unpredictable and uncontrollable planet-level existential threats. By moving beyond our only known home, we ensure species-level redundancy and guarantee that the light of consciousness is not tied to a single planet subject to the inevitable perils of a harsh and vast universe. We do not want humanity to suffer the Same fate as the dinosaurs. We wish to give them a reason to look to the future with excitement, to see that we are entering an age of abundance, with endless prosperity and an exciting future.
For decades, the reality of humans traveling between planets and stars has been captivating but remained locked in science fiction and screens. We have the ability to better understand the universe, to explore it, and ultimately to make life multiplanetary within it. We are becoming a civilization that can move beyond the cradle of Earth and begin inhabiting other worlds. While we remain committed to this fundamental mission, our progress in space access continues to yield opportunities to enrich life on Earth. For example, by drastically reducing the cost of reaching space, we have been able to expand our mission to address some of Earth’s most pressing challenges, including bridging the digital divide by aiming to connect the over 3 billion unconnected people to the internet and humanity's collective knowledge.
The rapid onset of the AI age has heightened the urgency of our mission, as AI holds the potential not only to accelerate space exploration but also to drive transformative societal progress on Earth. However, AI’s ability to transform human potential depends directly on meeting exponentially growing resource demands. On Earth, the massive expansion of data center capacity to support growing compute demand has significantly outpaced electricity generation, which has been largely flat in the U.S. for roughly 15 years, growing at a compound annual rate of only 0.1% from 2008 to 2023. Despite recent increases in AI Data Center power demand, annual U.S. electricity generation growth remained below 3% from 2023 to 2025, while China's electricity generation grew at roughly twice that rate over the same period. This supply-demand imbalance is already placing unsustainable pressure on terrestrial power grids, supply chains, and the environment. The sun contains approximately 99.8% of the solar system’s energy, and we therefore believe it is the only truly scalable solution to terrestrial energy constraints for the AI era. Harnessing this energy in space is far more efficient than on land. Space-based solar ARRays can generate over five times more energy per unit area than terrestrial solar due to continuous illumination, no atmospheric interference, and optimal orientation. SpaceX is well-positioned to capture this space-based solar energy, given our ability to rapidly access sun-synchronous orbit through our satellite manufacturing scale and launch capabilities. We are therefore expanding our footprint and harnessing the immense resources of space that are critical to sustaining technological progress. Our goal is to ensure AI becomes a force for human flourishing and civilizational benefit, rather than a catalyst for terrestrial resource depletion and instability.
We believe our current space efforts will catalyze transformative breakthroughs that could reshape terrestrial industries and lead to the emergence of new trillion-dollar markets on the Moon, Mars, and beyond. In particular, we believe that our goal of establishing a lunar presence will enable terawatt-scale annual AI compute growth, support deeper space exploration and industrialization, and serve as a stepping stone towards establishing a civilization on Mars. We believe that the next paradigm shift for humanity is the creation of a resilient, continuously expanding spacefaring civilization, driving sustained innovation on new frontiers and ultimately propelling us towards a Kardashev Type II civilization—we believe we can unlock an era of unprecedented economic expansion while simultaneously safeguarding humanity's future from existential risks.
Who We Are
SpaceX combines some of the most transformative and critical technologies in human history, including reusable rockets, global internet services, satellite-to-mobile communications, a real-time information, entertainment, and free speech platform, and a truth-seeking AI system designed to accelerate scientific discovery and augment human capability.
Our Unparalleled Launch Capabilities
Since its founding in 2002, SpaceX has cracked the code of large-scale space access, transforming an industry characterized by decades of stagnation, risk aversion, and economically distorted cost structures. We design, manufacture, launch, and refurbish reusable launch vehicles, offering cost-efficient, reliable, and high-frequency access to space for our own purposes and for third-party commercial and government customers. Our extensive vertical integration and end-to-end control over the entire value chain—from design to launch to operations—enable us to achieve unprecedented speed and cost efficiency. As of March 31, 2026, SpaceX had launched a total mass of approximately 7,400 metric tons to orbit via Falcon rockets, with a mission success rate exceeding 99%. We have completed approximately 650 orbital space launches, over 540 of which were conducted by flight-proven Falcon rockets. With the first successful launch of Falcon 1 in 2008, we became the first private company to successfully launch a liquid-fueled rocket into Earth orbit. In December 2015, we achieved what many considered impossible: landing a rocket back on Earth after it had launched into space. By 2017, we were routinely recovering and reusing Falcon 9 first-stage boosters, achieving another step-change reduction in the cost of space access through pioneering reusability. As of March 31, 2026, our Falcon 9 rocket had demonstrated the capability of flying a first stage 34 times. With the future deployment of Starship—designed to be the world’s first fully and rapidly reusable spacecraft—we aim to reduce the cost of reaching orbit by 99% or more compared to historical average launch costs, establishing the most affordable and scalable path to new opportunities in space, such as orbital AI compute and Mars exploration.
Our primary launch vehicles and spacecraft include:
Falcon 9. The world’s first orbital-class rapidly reusable rocket, Falcon 9 first launched in 2010 with a payload capacity of approximately 23 metric tons to low Earth orbit (LEO) in a fully expendable configuration. As of March 31, 2026, Falcon 9 had completed roughly 620 orbital space launches with a mission success rate exceeding 99%. According to NASA, the first version of Falcon 9 in 2010 reduced launch costs to approximately 18,500 per kilogram.
Falcon Heavy. Falcon Heavy first launched in 2018, sending a Tesla all-electric Roadster and its mannequin passenger, StArman, into orbit around the sun. With a payload capacity of approximately 64 metric tons to LEO, Falcon Heavy is a partially reusable super-heavy-lift launch vehicle designed to carry large payloads to orbit. Falcon Heavy is one of the most powerful operational rockets in the world by thrust at ignition and, as of March 31, 2026, had conducted 11 launches with a 100% mission success rate.
Dragon. Launched by Falcon 9 in 2012, our Dragon spacecraft became the first commercial spacecraft to deliver cargo to the International Space Station (ISS), an orbital laboratory serving as a research facility and destination for human spaceflight, and eight years later became the first privately built vehicle to fly humans to that orbiting laboratory. Since 2020, our Dragon spacecraft has safely transported 78 crew members from 20 nations.
Starship. First launched in 2023, Starship is designed to be a fully reusable super-heavy-lift launch vehicle. Starship V3 is designed to deliver 100 metric tons of payload to Earth orbit in a fully reusable configuration while achieving airline-like rapid turnaround times. Future generations of Starship are being designed to double its payload capacity. To date, we have executed 11 Starship flight tests. We have also planned a 12th flight test, which will debut a next-generation Starship vehicle and Super Heavy booster powered by the next-generation version of our Raptor engine and launched from a newly designed pad at Starbase. We expect Starship to begin delivering payloads to orbit in the second half of 2026. We have achieved innovation milestones, such as catching a booster with the "chopstick" arms at the same tower. We expect this capability to facilitate rapid refurbishment and reuse, allowing for multiple launches per day and lowering costs.
Having achieved rocket reusability, we recognized the immense potential of our launch business to create new revenue streams. This led to the development of Starlink, our global satellite internet constellation comprising thousands of LEO satellites, designed to deliver high-speed, low-latency broadband connectivity to severely underserved regions worldwide. While the concept of using satellites for global internet connectivity dates back decades, the technical challenges and prohibitive cost of accessing space and deploying the necessary satellites historically rendered such connectivity economically unviable. Within three years of our first satellite launch in 2019, we solved the technical and production challenges for the satellite, and within five years, we deployed the largest LEO constellation in existence. Today, Starlink is the only available low-latency network globally. By combining increasing launch cadence, expanding cargo capacity, and deCLIning unit costs driven by rapid reusability, we have generated compounding competitive advantages. This not only solidifies our core business but also provides expansive new market opportunities uniquely enabled by space.
Our Leading Capabilities Across Space, Connectivity, and AI
Space. While our launch capabilities support our other businesses, such as Starlink consumer broadband and Starlink Mobile, we also sell launch services to third-party customers. We offer launch services for satellite, cargo, and crewed missions to commercial, civil, international, and government customers via our reusable Falcon 9 and Falcon Heavy rockets. We are the primary launch provider for the U.S. government. In 2025, we executed 11 out of 12 National Security Space Launch (NSSL) medium and heavy missions and all five U.S. ISS crew and cargo missions for NASA.
Connectivity. Our Connectivity business includes Starlink Consumer Broadband, Enterprise Solutions, Government Solutions, and Starlink Mobile.
Starlink Consumer Broadband. We operate the world’s largest and most advanced space-based internet broadband service. We offer fiber-like download speeds—a median of 225 Mbps during peak hours for residential users as of March 31, 2026—and the technical capability to provide service anywhere on Earth, including the poles. This quality of service is supported by our massive network of approximately 9,600 LEO Starlink broadband and direct-to-mobile satellites, which as of March 31, 2026, accounted for roughly 75% of all active, maneuverable satellites in orbit. We expect to begin deploying next-generation V3 satellites, designed to provide 1 Tbps of downlink capacity per satellite, using Starship in the second half of 2026. We expect a single Starship launch will be able to deploy up to 60 V3 satellites to LEO, representing a 20-fold increase in Starlink downlink capacity deployment potential relative to a Falcon 9 launch.
Enterprise Solutions. SpaceX is a vital partner to a wide range of enterprises. We provide Starlink’s high-speed, low-latency, reliable internet service to enterprise customers in industries including construction, agriculture, retail, telecommunications, hospitality, aviation, maritime, and land mobility. Starlink’s unique capabilities are ideal for deployments at field offices, remote worksites, research stations, drilling platforms, rural hospitals, aircraft, cruise ships, trains, and hotels. We also serve a broad base of fixed-site customers in industries like retail and Financial Services that require high availability for critical operations and reliable connectivity in remote or hard-to-serve locations.
Government Solutions. For our government customers, we provide high-speed, resilient connectivity for public services, social impact, humanitarian efforts, and disaster response, even in the most remote and challenging environments. Starshield separately leverages our commercial LEO satellite constellation engineering and operational experience to develop a secure, dedicated satellite network designed specifically for U.S. government customers and national security applications.
Starlink Mobile. We provide satellite-to-mobile connectivity that supplements terrestrial networks and significantly reduces mobile dead zones in approximately 50 countries. Through partnerships with roughly 30 mobile network operators (MNOs) across six continents, we enable consumer, enterprise, and public sector customers to use existing cell phones in more places, support critical connectivity during disasters and blackouts, and unlock new applications for low-bandwidth mobile and IoT devices.
AI. We operate a highly vertically integrated AI platform.
AI Compute Infrastructure. xAI has established a leadership position in building and scaling ground-based AI compute infrastructure, becoming the first company to deploy a coherent gigawatt-scale AI training cluster. We own and operate what we believe to be the largest AI training data center cluster on Earth, including COLOSSUS and COLOSSUS II. The Teraflop chip manufacturing initiative, in Partnership with Tesla and Intel, aims to further extend our vertical integration into chip design and manufacturing to mitigate potential future chip shortages for SpaceX, optimize compute performance, and potentially reduce overall compute costs. In matters related to this collaboration, we have agreed with Tesla on a General Framework for Teraflop's future development. Any specific projects under this framework will be subject to separate negotiations and agreements, including any development timelines, milestones, and capital expenditures, and have not yet been determined. We believe the key constraint for continued AI growth is physical—chip manufacturing, data center infrastructure, and power generation; the future of AI will be determined by control of the physical stack.
Truth-Seeking Frontier Model. Since launching Grok-1 in November 2023, we have released four major versions and their significant variants, achieving one of the industry’s fastest iteration cycles. Within two years of the initial model release, Grok reached frontier-level performance in scientific reasoning (measured by GPQA Diamond score), on a timeline faster than reported by other leading model providers. Building on this trajectory, we expect to continue scaling Grok through subsequent generations. The ongoing training of the next-generation model is expected to scale toward trillions of parameters, which could represent a step-change in reasoning depth and overall intelligence. In this context, parameter count refers to the scale of the model, where parameters are numerical values within the model, such as "weights," adjusted during training to enable the model to recognize patterns and relationships in data. Larger parameter counts generally allow models to capture more complex relationships, store more knowledge, and achieve higher levels of reasoning capability. This accelerated innovation rate stems from our highly vertically integrated stack: full ownership of training infrastructure; access to the world’s most powerful compute clusters; and a relentless focus on truth-seeking and real-world utility. A key competitive differentiator is Grok's deep integration with X, giving it proprietary access to a real-time information stream of roughly 350 million daily posts, which enhances Grok’s freshness, relevance, and contextual awareness. This direct, real-time access to the information and human discourse on X strengthens Grok's truth-seeking capability by grounding its outputs in the latest knowledge and diverse perspectives.
Consumer and Enterprise Applications. We leverage our leading frontier models and compute infrastructure to deliver consumer and enterprise applications. We are also collaborating with Tesla on Macrohard, an Agentic AI platform designed to be able to fully simulate digital workflows and augment human computer operation using complex, Autonomous Agents. We believe Macrohard has the potential to fundamentally change how companies are structured and operate, thereby enabling a Dramatic increase in human Productivity.
Our Repeatable Business Model
Our business model is built on a repeatable, engineering-driven framework that combines our unparalleled launch capabilities, extreme vertical integration, rapid iteration, and disciplined capital investment to create durable, large-scale businesses. We execute this framework through the following core principles: leveraging our unparalleled launch capabilities to achieve scale; identifying and creating new trillion-dollar market opportunities; designing solutions with world-class engineering and first-principles thinking; applying "the algorithm" (reduce stupidity, delete, optimize, accelerate, automate); vertically integrating through to the end customer; continuously lowering costs and increasing throughput; and generating significant cash flow and reinvesting it into the future.
Our Engineering-First Culture
We achieve transformative technological breakthroughs because we accept only the laws of physics as limiting factors for our work and mission. Our core approach is deeply rooted in first-principles thinking, rejecting any preconceived notions or experience-based norms. We have a track record of achieving what many considered impossible. Our industry-defining accomplishments and historical milestones include being: the first private company to develop and launch a liquid-fueled rocket into orbit (2008); the first private company to successfully dock a private spacecraft with the International Space Station (2012); the first to successfully propulsively land (2015) and reuse an orbital-class rocket booster (2017); the first company to begin deploying a large-scale LEO broadband satellite constellation (2019); the first private company to transport astronauts to orbit, restoring U.S. capability to travel to and from the ISS (2020); the first to mass-manufacture consumer-grade phased-array user terminals (2022); the first to deploy a large-scale LEO satellite direct-to-mobile constellation (2025); the first to build a gigawatt-scale AI training cluster and the largest coherent supercomputer (2026); the first gigawatt-scale Megapack battery installation (2026); and the only company capable of building orbital AI compute at scale.
Our AI Compute Infrastructure Advantage and Growth Strategy
Why Compute Matters. We believe AI leadership will be determined by the ability to rapidly scale compute capacity to support exponential usage growth and frontier intelligence. The training and inference required by advanced AI models demand significant computational resources. Reasoning models introduced in 2024 demonstrated that allocating more compute resources and giving models more processing time during inference directly results in higher-quality intelligence. Furthermore, compute infrastructure that achieves end-to-end, cluster-level coherence through tight integration between software and hardware systems enables more efficient, stable, and higher-fidelity training and inference at scale—ultimately enhancing model intelligence and performance. In inference, we expect that computationally intensive reasoning, Agentic, and multimodal workloads will continue to grow as a portion of overall usage. Therefore, we believe operators with an advantage in model-to-compute integration (the compute power to effectively support and allocate training and inference workloads) are best positioned to win the AI race.
Self-Reinforcing Network Effects Between Lower Cost per Token, Model Quality, and User Adoption. AI systems are ultimately constrained or differentiated by the cost, speed, and scale at which they generate and process tokens. A "token" represents the fundamental unit of data consumed and produced by modern AI Models. This is because lower Cost Per Token enables more frequent model training, larger and more complex models, longer chains of reasoning and agentic workload processing, and significantly higher inference volumes at an economically viable price. This dynamic directly impacts model quality, responsiveness, and accessibility, while also determining the ability to serve growing global demand across consumer, enterprise, and mission-critical AI applications. This creates a self-reinforcing advantage where lower token costs drive higher model quality and user adoption, which in turn reinforces AI leadership.
Compute cost is the primary driver of cost per token. The total cost per token is dictated by the efficiency, availability, and unit economics of the underlying compute, as well as the cost of building and operating the compute infrastructure. Improvements in the cost of building and operating this compute infrastructure—whether through lower data center construction costs, lower power infrastructure costs, faster grid interconnection times, or higher cluster-level throughput—directly translate into lower cost per token. As a result, for a given level of intelligence, we expect the long-term economics of an AI company will be driven by the ability to consistently deliver frontier compute at the lowest possible cost per token. In simple terms, we view the cost per token as a function of three primary inputs—the underlying AI model, the compute hardware, and energy—and we expect to have competitive advantages in the latter two cost components. We believe we have a pathway to significantly reduce compute hardware costs over time through continued vertical integration and proprietary chip development, building on our experience designing custom Silicon for Starlink satellites. We also expect the marginal energy cost for our AI compute satellites to be very low, as our satellites are powered by space-based solar arrays. By driving the energy component towards minimal levels and pursuing improvements in compute hardware costs, we believe we can achieve meaningfully lower overall cost per token in the future.
We Have a Dual Speed and Cost Advantage in Ground-Based AI Compute. We own and operate what we believe to be the largest AI training data center cluster on Earth. Our AI compute facilities, COLOSSUS and COLOSSUS II, together provide approximately 1.0 gigawatt of compute power, with additional power capacity available for data center operations. Our first-principles thinking enables us to build coherent compute faster and at a lower cost than most of the rest of the industry. To bring compute clusters online as quickly as possible, we employ a vertically integrated, flexible construction approach. We brought the first cluster at COLOSSUS online in 122 days, repurposing the shell of an existing factory, and the first cluster at COLOSSUS II even faster, in 91 days. As an illustrative Comparison, the industry benchmark for bringing 100 megawatts of greenfield data center online is approximately two years. We have also demonstrated significant cost efficiency improvements, with data center construction costs at COLOSSUS II significantly below industry benchmarks on a per-megawatt basis.
We Believe Orbital AI Can Accelerate Time to Power and Reduce Token Costs. The sun contains approximately 99.8% of the solar system’s energy and presents what we believe is the only truly scalable solution to the challenge of accelerating compute demand relative to terrestrial energy constraints. The logical path forward is to shift energy-intensive AI workloads to orbit, where solar power is near-constant and uninterrupted. With this energy accessibility, we believe our launch operations will allow us to consistently activate the highest-performance hardware before competitors, reduce the time to useful tokens on frontier hardware, and sustain our token cost advantage. We believe SpaceX is in a unique position to deploy and operate data centers in orbit that could eventually achieve costs below terrestrial data centers over time, due to our extreme vertical integration approach across launch, scaled satellite manufacturing, network connectivity, and terrestrial data center expertise.
We Believe We Are Capable of Delivering Orbital AI Compute. We believe orbital AI compute is an incredibly hard technical problem that only we can solve at scale in the near term. We are the only company that has already completed the key technical challenges related to evolving a connectivity satellite into an AI compute satellite. We see ourselves as well-positioned to deliver a full-scale AI compute satellite constellation. Significant work remains, but we are confident in our unique leadership position. We possess the unparalleled satellite launch capability necessary to achieve large-scale deployment. Deploying 100 gigawatts annually via satellites with over 100 kilowatts of compute power per metric ton would require thousands of launches per year and the delivery of roughly 1 million metric tons to orbit annually. The fully reusable nature of Starship allows us to launch this level of mass. Starlink broadband V1 and V2 Mini satellites have demonstrated launch survivability and high reliability under vibration, shock, g-loads, acoustic stress, and vacuum exposure, achieving 99.9% average uptime. We have already solved many significant technical hurdles to evolving a connectivity satellite into an AI compute satellite. Through our leading expertise with connectivity satellites—including mass production, deployment, network operations, and inter-satellite laser and mesh connectivity—we have already solved the hardest parts of AI compute satellite development. Because an AI compute satellite represents an evolution of spacecraft engineering proven through Starlink, we believe the development of AI compute satellites is easier for us than for anyone else. Our existing Starlink constellation is another critical enabler for orbital AI compute, as its global network allows data from our AI compute satellites to reach ground stations anywhere on Earth. We will use our proven Starlink on-orbit technology to optimize our orbital AI compute. To operate orbital AI compute satellites, we plan to build upon our extensive experience operating roughly 9,600 LEO Starlink broadband and direct-to-mobile satellites. In 2023 alone, Starlink satellites autonomously performed over 1,000 collision avoidance maneuvers daily, guided by this technology, to operate the constellation safely and efficiently. This operational model allows us to control workload placement between Earth and space while maintaining Resilience through redundancy and fail-safe systems. High controllability will allow satellites to be optimized for brightness mitigation, disposal, and other operational modes. We can mass-manufacture our AI compute constellation with rapid upgrade cycles. We have built one of the world's largest satellite manufacturing operations. Our vertically integrated approach, with limited reliance on third-party suppliers, will be key to our scaling efforts and should enable us to deploy the latest AI processors. We believe SpaceX will be the first and only company capable of manufacturing satellites at automotive scale. We are building chip manufacturing capability to scale our access to AI compute hardware. In March 2026, we announced a collaboration with Tesla on the Terafab initiative, with a long-term goal of producing one terawatt of compute hardware annually. In matters related to this collaboration, we have agreed with Tesla on a general framework for Teraflop’s future development. Intel joined the project in April 2026 and is expected to contribute its expertise in designing, manufacturing, and packaging ultra-high-performance chips to help scale Terafab. Any specific projects under this framework will be subject to separate negotiations and agreements and have not yet been determined. With this in-house manufacturing capability, we plan to mitigate potential future chip shortages for SpaceX, especially as we develop orbital AI at scale, and design chips optimized for the space environment. We can leverage our terrestrial experience to build and operate compute clusters and AI workloads at scale. We believe our experience operating compute infrastructure on Earth provides the technical and operational foundation to extend these capabilities to orbit. For example, we plan to conduct extensive pre-deployment testing of compute hardware on Earth to identify early-life failures before launch, reducing on-orbit disruptions. For compute hardware that does experience failure, we plan to utilize existing Starlink fleet management software to reroute traffic to other satellites and prevent cluster-level downtime. We believe our infrastructure is uniquely positioned to deliver superior AI. We expect the combination of competitive cost per token, our ability to deploy and operate data centers in orbit, and our strength in connectivity will result in more scalable, globally accessible intelligence at high speed.
Our Strengths
Global leadership in orbital launch services.
Unmatched satellite and connectivity platform spanning design, manufacturing, deployment, and operations.
Truth-seeking AI model enhanced by real-time data.
Extreme vertical integration enabling high speed and superior cost efficiency at scale.
Unique capability to scale new trillion-dollar markets across Space, Connectivity, and AI.
A difficult-to-replicate business model.
A mission-driven culture and world-class talent.
Our Growth Strategies
Space:
Increase launch payload capacity.
Establish a lunar economy, including cargo transportation, manufacturing, and energy production on the Moon.
Connectivity:
Increase Starlink broadband users.
Expand our Starlink Mobile offerings.
Increase the capacity of our constellation.
AI:
Increase consumer AI platform monetization.
Increase X monetization.
Deepen enterprise and government adoption.
Increase the scale of terrestrial power and AI compute infrastructure.
Deploy orbital AI compute at scale.
Design and manufacture our own chips.
Launch Digital Human augmentation.
Future Markets:
Point-to-point terrestrial travel.
Space tourism.
In-orbit manufacturing.
Passenger and cargo transport to the Moon and Mars.
Energy production on the Moon and Mars.
Manufacturing capabilities on the Moon and Mars.
Asteroid mining.
Our Market Opportunity
We believe we have identified the largest actionable total addressable market ("TAM") in human history. We estimate our quantifiable TAM at 370 billion in the Space segment (from space-enabled solutions); 870 billion for Starlink Broadband and 26.5 trillion in the AI segment (760 billion for Consumer Subscriptions, 22.7 trillion for Enterprise Applications). To illustrate the scale of our addressable market opportunity, our global estimates exclude China and Russia.
Our Challenges
We face numerous challenges related to our business and growth strategies, ultimately tied to achieving our mission of making life multiplanetary, understanding the true nature of the universe, and extending the light of consciousness to the stars. Our mission-driven ethos guides our decision-making and forms the basis of our business plan, which premises on building, commercializing, and operating services and products at scales not previously realized. This objective requires us to develop and integrate complex, novel technologies, develop new processes and infrastructure, and coordinate across numerous suppliers, contractors, regulators, and stakeholders. Because we are attempting to execute at an unprecedented scale, we face significantly heightened uncertainty in design, engineering, procurement, construction, commissioning, and operational performance. In particular, our ability to execute our growth strategy is highly dependent on the successful development and scaling of Starship and our ability to increase launch cadence, both of which are subject to challenges and uncertainties inherent in the development and deployment of novel and complex technologies. Furthermore, many of the initiatives we describe under "Our Growth Strategies," including large-scale development of orbital AI compute, mass manufacturing of AI chips, establishing a lunar economy, transporting humans and cargo to the Moon and Mars, developing human augmentation systems, and more, involve significant technical complexity, unproven technologies, or technologies that do not yet exist, and these initiatives may not achieve commercial viability. Many of the innovative products and services described elsewhere in this prospectus may ultimately be unsuccessful and could require significant expenditures, innovations not yet achieved, or technologies not yet developed. As a result, timelines for certain initiatives involving unproven or new innovation, including our goals of deploying 100 gigawatts of annual Orbital Compute, establishing a lunar economy and interplanetary industrialization, and the launch cadence required to achieve these goals, may be difficult or impossible to determine. The execution of our growth strategy may take longer than anticipated, and you may not achieve a return on your investment within the expected timeframe, or at all. Furthermore, a portion of our expected market opportunity is associated with industries described under "Future Markets." Some of these industries, such as space tourism and lunar cargo transport, are still in their early stages. Other industries, including in-orbit manufacturing, lunar passenger and cargo transport, Mars passenger and cargo transport, lunar and Martian energy production, lunar and Martian manufacturing capabilities, and asteroid mining, do not currently exist. While we believe these industries will develop over time, the manner of their emergence, including the timing of commercialization, scale and pace of adoption, and the applicable competitive, technological, regulatory, geopolitical, and economic frameworks, could differ materially from our current expectations. Our Space, Connectivity, and AI segments also face the following challenges and uncertainties, among others:
Space: Our growth strategy depends on increasing launch cadence and payload capacity, which is contingent on the successful large-scale development of Starship. Unforeseen design modifications, supply chain disruptions, anomalies, environmental concerns, and other unforeseen technical challenges could cause delays or failures in Starship’s deployment timeline, thereby delaying or preventing us from achieving other business objectives, such as deploying next-generation satellites, expanding direct-to-mobile connectivity services, and deploying in-orbit AI compute infrastructure.
Connectivity: Our satellite connectivity, including global satellite-to-mobile connectivity services under Starlink Mobile, depends on access to radio frequency spectrum and authorizations from the U.S. Federal Communications Commission (FCC) and other national telecommunications regulators. Obtaining necessary authorizations can be a complex and time-consuming process. Without these licenses and approvals, we typically cannot provide connectivity services in a specific market. Spectrum access is inherently limited and highly regulated. Moreover, the growth of our connectivity services depends on increasing awareness and acceptance of Starlink connectivity in numerous international markets, each presenting unique challenges.
AI: Our AI business is in a relatively early stage, is being integrated into the organization, its Business Strategy is still being developed, and requires significant capital expenditure to fund compute, infrastructure and power generation, model training, and Product Development. Additionally, our AI business faces challenges inherent in a nascent, highly competitive, capital-intensive, and rapidly changing industry, including potentially disruptive technological changes, evolving industry and regulatory standards, emerging and well-funded competitors, frequent new product and service introductions, and shifting customer demands.
Any of the challenges described above, and others currently unknown to us, could negatively impact our business, financial condition, and results of operations. For a discussion of challenges, risks, and limitations that may impair our future prospects, please see elsewhere in this prospectus.
Recent Developments
Collaboration with Cursor. In April 2026, we entered into a Compute and Option Agreement with Anysphere, Inc. (doing business as Cursor, a San Francisco-based private software company), which we believe is a highly compelling extension of our vertically integrated compute infrastructure, model, and application strategy. Under the Compute Agreement, we will provide Cursor access to specified GPU cluster compute capacity and collaborate on improving existing models, including Grok, and potentially co-developing AI models and related model-specific deliverables or products. Under the Option Agreement, we have the right, but not the obligation, to acquire Cursor at a predetermined price or pay a fee. We view software development as a strategically important use case for AI because it combines high-quality, structured data, rapid feedback cycles, and frequent, mission-critical usage. AI-assisted coding workflows generate context-rich, verifiable data that can enhance model training and performance while driving sustained inference demand. Cursor’s deep integration with high-frequency coding workflows generates valuable developer interaction data, including code generation Prompts, iteration cycles, and Software Architecture decisions. We expect that access to this data will enhance our model training and inference, including for Grok. Simultaneously, by providing access to our large-scale compute infrastructure, we believe we can help Cursor deliver a faster, higher-quality user experience. The collaboration with Cursor may also accelerate our AI Strategy by more directly integrating our AI models into developer workflows and expanding the distribution of our AI capabilities through a high-engagement software interface. If a Cursor acquisition occurs following this offering, the acquisition consideration would consist of shares of our Class A common stock, based on an implied equity value for Cursor of 1.5 billion under the Option Agreement and deferred service fees of $8.5 billion under the Compute Agreement. These fees are payable in cash (or Class A common stock if this offering has not been completed at the time of fee payment). For more information on our arrangements with Cursor, including the option to acquire the company, see “Business—Collaboration with Cursor” elsewhere in this prospectus.
Compute Services Agreements with Third Parties. We believe our compute infrastructure and related strategy provide robust flexibility in how we allocate and monetize capacity. We are able to use compute resources to support our proprietary AI applications, such as Grok 5, currently training on COLOSSUS II, while also offering selected compute capacity access to third-party customers. For example, in May 2026, we entered into Cloud Services Agreements with anthropic PBC, an AI research and development public benefit company, for access to compute capacity at COLOSSUS and COLOSSUS II. Pursuant to these agreements, the customer agreed to pay us $1.25 billion per month through May 2029, with capacity ramped at lower fees in May and June 2026. Either party may terminate upon 90 days' prior notice. The customer retains ownership of and intellectual property rights to its content, AI models, and related data. This structure allows us to monetize unused compute capacity in our infrastructure while still allowing for capacity to be reallocated back to our own internal initiatives should the need arise in the future. We have sufficient capacity to power our own AI models, including support for training and inference demand, and to fulfill our obligations under these agreements. We expect to enter into more similar service contracts. We believe this opportunity highlights the growing significance of large-scale, frontier-grade AI Infrastructure and positions us as a differentiated provider of high-performance compute capacity for both internal and third-party AI workloads. We believe our dual monetization strategy provides multiple paths to returns on invested capital.
Founder, Chief Executive Officer, Chief Technical Officer and Chairman of Our Board
Mr. Musk is our founder, Chief Executive Officer, Chief Technical Officer, and Chairman of our Board. Assuming an initial public offering price of $___ per share (the mIDPoint of the estimated price range set forth on the cover page of this prospectus) and the number of shares offered as set forth on the cover page of this prospectus, Mr. Musk will own approximately ___% of the voting power of our common stock following this offering (or approximately ___% if the underwriters exercise their over-allotment option in full) through his holdings of our Class A common stock and Class B common stock. Under our charter, the holders of Class B common stock are entitled to elect a majority of the Board (such directors, the “Class B Directors”) so long as any shares of Class B common stock remain ouTSTanding. As the holder of a majority of the shares of Class B common stock, Mr. Musk will be able to elect, remove, or fill any vacancy among the Class B Directors. Additionally, so long as he beneficially owns more than 50% of the voting power of our common stock, Mr. Musk will control the vote for the election of our directors. Accordingly, Mr. Musk will have the power to control the outcome of matters requiring stockholder approval, including the election of all of our directors, and control our business and affairs.
Our Controlled Company Status
Following this offering, we will be a controlled company under Nasdaq and Nasdaq Texas listing rules. Controlled companies are not required to have a board of directors composed of a majority of independent directors, or to establish independent compensation and nominating committees. As a controlled company, we remain subject to rules requiring us to have an audit committee composed entirely of independent directors.
Corporate Information
We were incorporated on March 14, 2002, as Space Exploration Technologies Corp., a Delaware corporation, and reincorporated on February 14, 2024, as a Texas corporation. Our principal executive offices are located at 1 Rocket Road, Starbase, Texas 78521. Our website address is www.spacex.com. Information contained on, linked from, or otherwise connected to our website does not constitute part of this prospectus or the registration statement of which it forms a part and is not incorporated by reference herein.
Summary of Risk Factors
An investment in our Class A common stock involves risks and uncertainties. The following is a summary of the principal factors that make an investment in our Class A common stock speculative or risky, all of which are more fully described in the “Risk Factors” section below. This summary should be read in conjunction with the “Risk Factors” section and is not intended to be an exhaustive summary.
Failure or delay in the large-scale development of Starship, or the subsequent inability to achieve the required launch cadence, reusability, and capability, will delay or limit our ability to execute our growth strategy, including deploying next-generation satellites, global satellite-to-mobile connectivity, and orbital AI compute, which could have a material adverse effect on our business, financial condition, results of operations, and future prospects.
Any delay or difficulty in obtaining, maintaining, or renewing the regulatory approvals and licenses required for our space-related activities, including Federal Aviation Administration (FAA) launch and re-entry licenses, would significantly delay or disrupt our operations, harm our business, or limit our ability to execute our business strategy.
Any delay or difficulty in obtaining, maintaining, or renewing the communications licenses and spectrum authorizations required for our satellite connectivity services, including international and FCC satellite spectrum licenses, would significantly delay or disrupt our operations, harm our business, or limit our ability to execute our business strategy.
Our AI products and X platform are subject to complex and evolving U.S. and foreign laws and regulations that may change and be subject to uncertain interpretation, and we may be required to make modifications to our products and business practices and face fines, increased operating costs, declines in user growth or engagement, customer loss, or other harm to our AI products and the X platform.
Our business strategy depends on the successful design, development, and deployment of our products and services and related platforms, infrastructure, and other strategic initiatives at an unprecedented scale, which presents significant execution, cost, and timing risks.
We have experienced and will likely continue to experience launch delays and failures, which could have a material adverse effect on our business, financial condition, results of operations, and future prospects.
Our satellites, launch vehicles, and other space-based technologies operate, and in the case of orbital AI compute will operate, in a harsh and unpredictable space environment, exposing them to a wide range of unique space-related risks that could cause their malfunction or failure, any of which could adversely affect our business, financial condition, results of operations, and future prospects.
The continued proliferation of low Earth orbit satellite constellations and the risk of collision with space debris or other spacecraft could limit or impair our launch flexibility and satellite deployments, potentially adversely affecting our business, financial condition, results of operations, and future prospects.
Disruptions to critical satellite network, ground station, launch, manufacturing, or spacecraft or data center infrastructure operations could result in significant downtime, operational delays, or service loss, any of which could materially and adversely affect our business, financial condition, results of operations, and future prospects.
The manufacturing, testing, and launching of rockets, satellites, and spacecraft, including our efforts to reuse rockets and spacecraft, involves inherent risks that could result in Personal injury or death, property damage, and environmental harm or other adverse environmental impacts. Any such event could result in significant losses, including reputational harm and legal liability, which could materially and adversely affect our business.
While we focus on vertical integration of our business, we rely on third parties to manufacture and supply certain key components necessary for our launch, connectivity, and AI Services, and any supply shortages or disruptions, or failures in their performance, could have a material adverse effect on our business, financial condition, results of operations, and future prospects.
Comments & Questions (0)
No comments yet
Be the first to comment!