SpaceX just dropped nearly 400 pages of financial detail onto the SEC, and buried in that filing is one of the most audacious statements a company could possibly make: “We believe we have identified the largest TAM in human history.” That’s total addressable market, for those keeping score. The number? $28.5 trillion.
Let that sink for a moment. That’s not a typo. That’s SpaceX, a company that built its reputation launching rockets into space, claiming it can reach a market larger than global GDP.
The filing comes ahead of an anticipated IPO in June, and after nearly 25 years as a private company, it offers the first comprehensive look at SpaceX’s financials. The picture is both impressive and deeply strange.
The Money Story
SpaceX pulled in $18.67 billion in revenue last year, up from $14.02 billion in 2024. That’s legitimately solid growth. But here’s where things get interesting: the company lost $4.94 billion in 2025. After turning a small profit the year before, SpaceX swung hard into the red. The culprit? Spending on artificial intelligence development.
This is crucial context. SpaceX isn’t just talking about AI as some future play. It’s actively hemorrhaging cash to build it today. Elon Musk takes home $54,080 annually, pegged to California’s minimum salary for exempt employees, a move that feels more performative than anything else. Gwynne Shotwell, the president and COO, earned $85.8 million in total compensation, including stock awards.
The launch business itself remains relatively opaque in the filing. We know the Falcon 9 costs roughly $15 million to launch but sells for $74 million publicly, yet SpaceX doesn’t break this down officially. With Starship, the company aims for $185 per kilogram to orbit, which would be a genuine game-changer for heavy lift if achieved.
The AI Gamble
Here’s where SpaceX makes its real claim: out of that $28.5 trillion TAM, only about $2 trillion relates directly to space or Starlink. The remaining $26.5 trillion? That’s supposedly AI compute.
The company bases these estimates on data center projections from sources like the RAND Corporation, combined with internal assumptions about global compute capacity, power usage, and pricing. SpaceX doesn’t soft-pedal its vision either. It claims to be “the only company with a commercially viable path to building orbital AI compute at scale” and targets launching 100 gigawatts of compute to space annually over time.
Deployment of these orbital AI compute satellites could begin as early as 2028.
The logic is straightforward: SpaceX has unmatched launch capability and can manufacture satellites at scale. Therefore, it’s best positioned to build a constellation of orbital data centers. Whether investors buy this depends largely on whether they believe orbital compute makes economic sense at the scale SpaceX imagines. That’s not a solved problem yet.
The Control Question and Political Risk
After the IPO closes, Musk will retain 85.1 percent of combined voting power. He’ll serve as CEO and board chair, and removing him would be practically impossible. For better or worse, this company’s future is Musk’s future.
The filing also contains a passage about political risk that feels almost prophetic. SpaceX notes that shifts in Congress or the presidency could “result in significant changes in government spending priorities, regulatory posture, and the allocation of contracts and resources.” The company acknowledges that Musk’s status as an advisor to President Trump means government relationships could swing dramatically depending on which party holds power.
That’s not theoretical concern. SpaceX already operates the vast majority of US launch capacity and manages critical national security contracts. Government favor matters immensely to the bottom line.
The Scale Problem
SpaceX has come genuinely far. It launches roughly 80 percent of all mass put into orbit annually and operates more satellites than the rest of the world combined. Those aren’t small accomplishments. The company went from launching a dinky Falcon 1 to the world’s most accomplished launch operation in two decades.
But reaching a $28.5 trillion valuation requires transformation from space company to AI company. That’s not evolution. That’s a fundamental business model shift. The company acknowledges it still has enormous work ahead on Starship to make it fully reusable and capable of Moon and Mars missions. Those challenges are “novel or untested” and require “substantial capital investment,” according to the filing itself.
So SpaceX is simultaneously trying to perfect reusable super-heavy lift, maintain launch dominance, operate the world’s largest satellite constellation, and build a new orbital AI compute infrastructure. That’s not ambition. That’s a lot of plates spinning at once.
The question isn’t whether SpaceX can execute some of these things. It’s whether it can execute all of them while the markets decide whether any of this is actually worth the astronomical valuations being discussed. The filing gives us the company’s answers. Whether those answers are right depends on execution across fronts we haven’t fully tested yet.
