Space-X: An Astronomical IPO

After months of speculation, Space Exploration Technologies Corp. (SpaceX) has announced its intent to go public on the Nasdaq on June 12th. At a targeted $1.75 trillion valuation and with up to $75 billion being raised, it will be the largest IPO in history, roughly three times the size of Saudi Aramco’s 2019 offering. The accompanying Form S-1 offers a glimpse into one of the most ambitious companies ever built.

Three segments sit under one roof: Space, Connectivity, and AI. The Space business has already rewritten the economics of an industry, with reusable rockets reducing the cost per kilogram to orbit by roughly 85%. SpaceX now accounts for more than 80% of global launch weight, and has made once-speculative concepts like orbital data centres economically plausible. That launch infrastructure also underpins Starlink’s formidable position. As a subsidiary, Starlink pays no launch toll, giving it a structural advantage no competitor can replicate. With 10.3 million subscribers across 164 countries and a 39% operating margin, the Connectivity segment generates a dominant 61% of SpaceX’s overall revenue.

Yet the prospectus also reveals a striking contradiction between SpaceX’s interplanetary messaging and its near-term economic reality. While the company details Mars colonies and orbital civilisation, its self-defined $28.5 trillion Total Addressable Market is overwhelmingly terrestrial. Nearly 80% of that opportunity sits within enterprise AI, meaning the bull case depends less on colonising Mars, and more on monetising compute capacity and automation software. Funding this buildout, however, is enormously expensive. In the first quarter of 2026 alone, SpaceX posted a free cash burn of $9.1 billion, driven by $10.1 billion in capital expenditures. Remarkably, the AI segment accounted for just 17% of revenue but devoured 87% of total capital expenditure, effectively subsidised by Starlink’s satellite cash flows. However, their AI segment does have some optionality, as whatever data centre capacity they don’t fully utilise can be quickly rented out to their compute-constrained competitors. In that vein, Anthropic has reportedly agreed to pay $15 billion annually through 2029 for access to its Colossus compute infrastructure, a contract large enough to materially reshape the segment’s economics.

To shield this aggressive capital allocation, SpaceX is deploying one of the most founder-centric governance structures ever brought to public markets. Dual-class shareholding means Musk will retain 85.1% voting control, allowing him to bypass independent board requirements, pick his own compensation committee, and secure the freedom to compete directly against his own public entity. Texas business court structures also virtually insulate management from shareholder scrutiny. To even file a derivative lawsuit for breach of fiduciary duty, an investor must own at least 3% of the company, roughly $50 billion worth of shares. Whether this concentration of power proves visionary or dangerous will depend entirely on one’s faith in Musk himself. His own payout is tied to perhaps the most audacious promise in corporate history: the establishment of a permanent human colony on Mars containing at least one million inhabitants. At a valuation of $1.75 trillion, investors are not merely underwriting current economics, but decades of execution across industries that do not yet fully exist.