The New Development: Experts Are Now Talking

Our previous coverage broke down the S-1's raw numbers — $20 billion into AI in one year, $37 billion lost since inception, a $2 trillion price tag on the orbital satellite vision. What's changed: the expert community is now responding publicly, and the response has been critical.

Robert Zubrin — aerospace engineer, rocket designer, and founder of the Mars Society — told Forbes contributor Kevin Holden Platt that Musk's plan to launch one million AI data center satellites starting in 2028 could trigger a "financial catastrophe" and send SpaceX into a "high-speed nosedive."

Zubrin isn't a think-tank blogger. He's one of the most credentialed rocket engineers alive.

The Core Math Problem

Zubrin's argument is straightforward. SpaceX has already launched roughly 10,000 Starlink broadband satellites. Each one cost approximately $2 million to build and launch, according to Zubrin's estimate cited in Forbes.

Scale that to one million satellites using the same assembly and launch economics: $2 trillion.

That's the entire projected valuation of SpaceX at IPO. The company would be spending its entire market cap — and then some — just to build the constellation. Before it generates a single dollar from it.

Musk's counter-argument is that Starship changes the economics entirely. He's promising Starship launches once every hour to get these satellites up and predicts the orbital AI network becomes commercially viable "within 2 to 3 years" of deployment.

Starship is still experimental, and these remain projections rather than established timelines.

What the S-1 Actually Says vs. What Musk Is Promising

The S-1 filing — made public May 20 according to TechCrunch — shows SpaceX's AI-related spending consumed roughly 60% of all capital spending in 2025, or around $20 billion. That division lost billions and grew revenue by only about 22%, according to TechCrunch.

For context, TechCrunch noted that 22% growth is "far below the reported revenue growth rates at frontier AI labs." The AI division is currently operating at a loss relative to its spending.

Starlink, meanwhile, is the only thing keeping the lights on. It generated over $11 billion — more than half of SpaceX's $18 billion in 2025 revenue, per the S-1 as reported by TechCrunch and NPR.

The company lost $4.9 billion last year on that $18 billion in revenue, as Reuters first reported and the S-1 confirmed.

The most successful commercial rocket company in history is burning nearly $5 billion a year — and is now proposing a project that costs 400 times that annual loss figure.

The IPO Numbers Keep Shifting

When the S-1 first dropped, early reports pegged the fundraise target at around $75 billion. According to NPR, that number has since been reported as high as $80 billion, which would make it the largest IPO in history — dwarfing Saudi Aramco's $29 billion raise in 2019.

The valuation range is also still in flux. TechCrunch reported a figure of approximately $1.75 trillion. NPR said the company could be valued at "over $1 trillion" and potentially more than Tesla. The mid-June target date for the actual listing is what's now on the clock.

SpaceX has chosen the ticker SPCX on the Nasdaq.

What Mainstream Coverage Is Missing

Most of the coverage — from NPR to TechCrunch — is treating this as a triumphant tech story. Musk goes public, enriches himself further (he controls 85% of voting power, per the S-1 cited by NPR), Wall Street celebrates.

But the orbital data center ambition isn't just a moonshot. It's a plan that, by Zubrin's analysis, would require spending the company's entire market value on infrastructure before the revenue model is proven. Terrestrial data centers are already cheaper. The fundamental economic case for putting AI compute in orbit — versus on the ground — remains unproven.

Also underreported: Musk holds CEO, CTO, and chairman of the board simultaneously, according to TechCrunch. That's an unusual concentration of control for a publicly traded company. It means shareholders will have virtually no ability to challenge his direction, even if the orbital data center bet starts looking like a catastrophe.

The S-1 also discloses that SpaceX is on the hook for legal battles tied to Musk's other companies — specifically xAI and social media ventures — that are expected to cost $530 million, per TechCrunch. That's not SpaceX's core responsibility. But SpaceX shareholders will pay for it.

What This Means for Regular People

If you're thinking about buying SPCX when it lists: understand what you're actually buying. You're buying a Starlink cash cow attached to an experimental rocket program attached to an underperforming AI division attached to a $2 trillion orbital plan that the company's own pre-IPO valuation can't cover.

Zubrin's warning deserves broader attention than a Forbes contributor piece. The experts are talking. Wall Street's response will come in mid-June.