Letter 85 asked why humanity builds in space at all. This letter answers a different question: what is it actually worth? The global space economy reached $626 billion in 2025 — larger than the GDP of Sweden, Belgium, or Poland. Commercial activity now drives 78% of that total, a complete inversion from a decade ago. Starlink alone generated $11.39 billion in 2025 and crossed 10 million subscribers across 164 countries, growing faster than any subscription business in telecom history. Launch costs have fallen 90% in two decades. This is no longer a frontier. It is an industry — measurable, investable, and astonishingly large.
Not investment advice. Data sourced from Orbital Radar Space Economy Tracker June 2026, Novaspace/SpaceNews 2025 estimates, Sacra SpaceX Financial Analysis 2026, GM Insights Space Economy Market Report April 2026, BryceTech, Sci-Tech Today Commercial Space Statistics. All figures current as of June 2026.
Letter 85 made the case that humanity builds beyond Earth not from necessity but from wonder — that the most honest investment thesis in space is also a philosophical conviction about what humanity is for. That remains true. But conviction alone does not move $626 billion of annual economic activity, and the space economy of 2026 has crossed decisively from philosophical aspiration into measurable industrial reality. The global space economy reached $626 billion in 2025 — comparable to the entire GDP of Poland, larger than Sweden or Belgium — and every major forecaster from BryceTech to Novaspace expects it to exceed $1 trillion by the early-to-mid 2030s.
The most consequential shift in the data is not the absolute size but the composition. Commercial activity — satellite services, broadband, launch services sold to private customers — now represents approximately 78% of the global space economy, up from roughly 60% a decade ago. Government spending, which once dominated every dollar spent beyond the atmosphere, has been structurally overtaken by private commerce. This is the same transition that the internet underwent in the 1990s — from a government and military research project to a commercial infrastructure layer that the private sector built, owns, and monetises at a scale the original government sponsors never imagined.
If there is one number that explains why space went from a government prestige project to a $626 billion commercial industry in under two decades, it is the cost of getting a kilogram of payload into orbit. Every business model in this letter — broadband constellations, Earth observation, in-space manufacturing, lunar logistics — depends on this single cost curve continuing to fall. It has fallen further and faster than almost any other industrial input in modern economic history.
The 90% cost collapse that made the commercial space economy possible
Satellite manufacturing costs have collapsed in parallel — from $100,000 per kilogram for traditional bespoke satellites to approximately $1,000 per kilogram for Starlink's mass-produced units. SpaceX achieved this cost advantage through 5–10× lower cost per ton than competitors, vertically integrating 70% of Falcon 9 production in-house and replacing industry-standard components with commercial-grade alternatives — a $100,000 industry-standard radio replaced with a $5,000 commercial equivalent, for example. The Falcon 9 booster has completed over 260 successful re-flights out of 315 total launches, amortising the roughly $62 million launch cost across multiple missions while maintaining a success rate above 99%.
The largest single segment. Broadband connectivity, television and video distribution, navigation, and Earth observation data sold as a service. The declining satellite video broadcasting segment still represents some $72 billion in legacy revenue even as growth shifts decisively toward broadband.
The receivers, terminals, antennas, and infrastructure that turn satellite signals into usable connectivity on Earth. Often the largest segment by revenue yet the least discussed — every Starlink dish, every GPS chip, every satellite TV receiver sold counts here.
The smallest segment by revenue but the foundation everything else depends on. SpaceX commands over 60% market share by launch count globally and roughly 85% of all US orbital launches — an effective pricing ceiling that every competitor, from Rocket Lab to Blue Origin to China's CASC, must operate beneath.
Building the satellites themselves. Smallsat production runs $5–20 million per unit for LEO constellation satellites; traditional geostationary satellites still cost $200–300 million per unit. Starlink's mass-production approach — over 4,000 satellites per year at peak capacity — has no precedent in the history of any single spacecraft product line.
If the space economy has a single proof point demonstrating that commercial space is genuinely, durably profitable rather than venture-subsidised speculation, it is Starlink. The growth trajectory has no comparable precedent in the history of telecommunications: 10,000 beta users in 2021, 1 million by 2022, 2.3 million by end-2023, 4.6 million by end-2024, over 9 million by end-2025, and surpassing 10 million active customers by February 2026 across 160-plus countries and territories. Revenue reached $11.39 billion in 2025 — representing 61% of SpaceX's total revenue — with analysts projecting approximately $20 billion in revenue and $14 billion in EBITDA for 2026.
The business model evolution is instructive. SpaceX's draft IPO prospectus disclosed that average revenue per subscriber actually fell 18% to $81 per month between 2023 and 2025, even as the subscriber base quadrupled — a deliberate strategic trade of ARPU for global volume during the land-grab phase of constellation deployment. In May 2026, SpaceX reversed that trend, raising Starlink plan prices by up to $10 per month — signalling a decisive shift toward monetising its now-massive installed base rather than continuing to prioritise pure subscriber growth. This is the same playbook every dominant network platform follows once scale is achieved: land-grab first, monetise second. Starlink now represents approximately 75% of all active, manoeuvrable satellites in Earth orbit, with manufacturing capacity exceeding 4,000 satellites annually — competitors including Amazon's Kuiper (roughly 200 operational satellites by 2026 despite a committed $10+ billion investment) remain meaningfully behind.
This Terafab ambition connects directly to the semiconductor reindustrialisation thesis explored in Letter 97 — SpaceX is attempting to vertically integrate not just rockets and satellites but the chips that power its AI-enabled satellite constellation, in direct parallel with the broader Western push toward semiconductor self-sufficiency. The capital intensity is enormous and the risk is real: SpaceX's draft prospectus also disclosed a nearly $5 billion GAAP loss in 2025, tied substantially to costs absorbed from the xAI merger, alongside approximately $17 billion in combined space and AI cash burn. Elon Musk has himself warned of "genuine risk of bankruptcy" if Starship fails to achieve a flight cadence of at least once every two weeks — a reminder that even the dominant company in the most successful space business in history operates with genuine execution risk at this scale of ambition.
The most credible challenger to SpaceX's launch dominance, pursuing a strategy of mass-produced reusable small launchers targeting sub-$2,000/kg costs. Rocket Lab's Neutron scale-up is the leading indicator the industry's disruption analysts are watching most closely — if it achieves its cost targets on schedule, it represents the clearest evidence that SpaceX's launch cost advantage is replicable by a well-executed competitor rather than a permanent monopoly position.
Amazon's most direct competitive answer to Starlink, leveraging the company's logistics, cloud infrastructure, and balance sheet scale. The gap remains enormous in practice — roughly 200 operational satellites against Starlink's 10,000-plus — demonstrating just how difficult Starlink's first-mover manufacturing and launch cadence advantage is to close even for a company with Amazon's resources. Initial launches are scaling through 2025–2026 aboard Atlas V, Vulcan, and New Glenn vehicles.
The traditional European space primes, growing steadily through satellite communications, military applications, and strategic partnerships like Project Bromo. Revenue growth at 11% for both companies reflects healthy demand for traditional government and defence-anchored space programmes — a reminder that even as commercial space captures headlines, the established primes serving sovereign and military customers remain durable, profitable businesses with long-term contracted revenue.
The leading commercial space station developer, positioning for the post-International Space Station era. Commercial space station funding is expected to surpass $20 billion by 2030 as microgravity research, pharmaceutical development, space tourism, and industrial manufacturing applications mature. Axiom's first module launch — targeted for 2026 — represents the first real test of whether a fully commercial orbital habitat business model is viable independent of government anchor tenancy.
The launch services market — the most visible, most discussed segment of the space economy — is also the smallest by revenue, at roughly $14–18 billion annually against a $626 billion total industry. The actual economic value of space is overwhelmingly downstream: ground equipment, satellite services, and the data and connectivity applications that satellites enable. Investors and commentators who focus exclusively on rockets and launch providers are analysing the most exciting but least economically significant layer of the stack. The genuinely large economic opportunities are in Earth observation analytics (growing at roughly 15% annually across agriculture, climate tracking, insurance, infrastructure, and logistics applications), ground equipment manufacturing, and the LEO broadband market specifically, projected to grow from roughly $9 billion in 2025 to $47–55 billion by 2034.
SpaceX's dominance is so complete that it creates genuine systemic concentration risk for the entire commercial space economy. A company controlling 85% of US orbital launches, 75% of all active manoeuvrable satellites in orbit, and the dominant share of global commercial launch revenue represents the kind of concentration that, in any other infrastructure industry, would attract serious antitrust scrutiny. SpaceX's own disclosed near-$5 billion GAAP loss in 2025 and Musk's explicit warning about bankruptcy risk if Starship cadence targets are missed is a reminder that the entire commercial space economy's near-term trajectory is more dependent on a single company's execution than almost any other major industry analysed in this letter series.
The space economy's growth projections — $1 trillion by the early-to-mid 2030s, $2 trillion by 2040 in the most aggressive estimates — assume that the launch cost curve continues falling and that LEO broadband demand keeps compounding at current rates. Both assumptions are plausible but not guaranteed. Starship's flight cadence challenges, the genuine technical difficulty of sustained Mars-capable reusability, and the possibility that LEO broadband subscriber growth slows as the addressable underserved population is saturated are all real risks to the consensus bull case. The space economy's bull case is well-evidenced by the data in this letter. It is not risk-free, and the concentration in a small number of companies — chiefly one — makes it more fragile than the trillion-dollar headline numbers suggest.
Long-horizon thinking on capital, technology, and the forces shaping the next decade of wealth creation. Written from first principles. Not consensus. Not noise.