SpaceX has filed to go public, and the number that should be making the rounds today is $1.75 trillion.
That's the valuation the company is targeting in its confidential IPO filing, according to CNBC, with a listing expected around June and a target raise of up to $75 billion. Both figures would shatter records. The previous largest IPO was Saudi Aramco at $29.4 billion in 2019. SpaceX is trying to raise more than double that in a single offering. The company filed the paperwork on April 1, SpaceNews reported.
Before the market can decide whether to write that check, it has to reckon with what SpaceX actually is today: and what it is today is Starlink.
The satellite internet business generates between 50 and 80 percent of SpaceX's revenue, Reuters reported, with more than 9,500 satellites deployed since 2019 and over 9 million subscribers globally. SpaceX generated roughly $8 billion in profit on $15 billion to $16 billion in revenue in 2025, following $13.1 billion in 2024 revenue that was itself a 51 percent jump from the prior year, Reuters reported. Shay Boloor, an analyst at Griffin Securities, told Reuters that Starlink is the only reason the $1.75 trillion valuation is defensible. "This is going to be the recurring revenue engine with 9 million subscribers, defense contracts and its own data network," he said.
That revenue engine is also the product of a significant federal subsidy tailwind. SpaceX has received more than $24.4 billion from the federal government since 2008, CNBC reported, citing FedScout data. The company's commercial launch business, 165 orbital flights in 2025, is real, but it's not what justifies a $1.75 trillion multiple. That multiple is a bet on Starlink's subscriber curve continuing to bend upward, on defense contracts expanding, and on a data network that hasn't fully materialized yet.
SpaceX is also asking public-market investors to price in a merger that closed two months ago. SpaceX merged with xAI, Elon Musk's artificial intelligence startup, in February 2026 at a combined valuation of $1.25 trillion: $1 trillion for the rocket company and $250 billion for the developer of the Grok chatbot, Reuters reported. xAI posted a net loss of $1.46 billion in the September quarter on just $107 million in revenue, Reuters reported, meaning the AI business is currently a cost center, not a revenue line. The IPO prospectus presumably lays out how those financials are meant to combine.
Then there's the part of the pitch that sounds like it was written for a different century.
SpaceX filed an FCC application in late January for up to 1 million satellites in an orbital AI data center constellation, Data Center Dynamics reported. The satellites would use radiative cooling and are designed for a five-year operational lifespan. The ambition is to put compute in orbit, away from terrestrial power and land constraints, and the filing suggests IPO proceeds would fund it.
The physics of that plan are not kind. MoffettNathanson estimated that achieving Musk's stated goal would require 3,000 Starship launches per year, or roughly eight per day, Reuters reported. Launch costs would need to fall from thousands of dollars per kilogram today to the low hundreds. The total price tag for the orbital data center plan would run into the trillions, MoffettNathanson calculated. Nvidia CEO Jensen Huang, asked about the economics of space-based AI infrastructure, was direct: "We should definitely work on the ground first because we are already here," Reuters reported.
SpaceX will host an analyst day on April 21, with an optional tour of the xAI Macrohard data center in Memphis on April 23, two days later. The schedule suggests the company expects Wall Street to do its homework on both businesses simultaneously.
Some of that homework will be easier than others. Starlink's subscriber count is a number. The defense contract pipeline has documents. Starship's flight test cadence, five tests in 2025 with the first three ending in explosive setbacks and the last two successful, Reuters reported, is a matter of public record. The orbital data center plan requires a launch cost reduction that hasn't happened yet, a manufacturing scale that doesn't exist yet, and a Starship flight rate that is eight launches per day away from being real. That part of the pitch is asking investors to buy a rocket that hasn't finished its test program and a satellite constellation that doesn't have a viable path to deployment yet.
Armand Musey, a space industry analyst, told SpaceNews that the IPO is the most anticipated in history and that the valuation questions remain unresolved. "It can't," Musey said of the current company's ability to justify the price talk.
SpaceX has shipped what it promised. The question the IPO is asking is whether the next version of that promise is worth $1.75 trillion before the first one has finished loading.
† Revenue grew 51% to $13.1 billion in 2024, according to SpaceX.†
†† For 2025, SpaceX generated approximately $15 billion to $16 billion in revenue with roughly $8 billion in EBITDA profit, according to SpaceX.†
††† serves more than 9 million users globally across a constellation of roughly 9,500 satellites, per SpaceX.†
†††† generates between 50% and 80% of SpaceX's total revenue, per SpaceX.†
† Revenue grew 51% to $13.1 billion in 2024, according to SpaceX.†
†† For 2025, SpaceX generated approximately $15 billion to $16 billion in revenue with roughly $8 billion in EBITDA profit, according to SpaceX.†
††† serves more than 9 million users globally across a constellation of roughly 9,500 satellites, per SpaceX.†
†††† generates between 50% and 80% of SpaceX's total revenue, per SpaceX.†
