What's New Since Our Last Coverage

The S-1 is no longer confidential. SpaceX filed publicly with the SEC on May 20, 2026. Since our last article covered the roadshow and initial market reaction, the focus has shifted to what's inside the filing and what serious critics are saying about the plan.

The AI Spending Number

According to TechCrunch, SpaceX directed 60% of its 2025 capital expenditures to its AI division — approximately $20 billion in a single year. That AI division, which houses the Grok chatbot inherited through the xAI merger, lost billions and grew revenue by only 22%.

For context: frontier AI labs like Anthropic and OpenAI are posting revenue growth rates well ahead of that. SpaceX's AI division isn't keeping pace, and it's burning cash at a rate that would concern most CFOs.

The S-1 also reveals SpaceX has lost more than $37 billion since inception, and posted a $4.9 billion loss in 2025 on revenues of more than $18 billion, per Reuters reporting cited in the TechCrunch analysis.

Starlink is carrying the entire company right now — generating over $11 billion of that $18 billion in revenue, according to the filing.

Musk Controls the Whole Thing

NPR reported that Musk holds 85% of voting power in SpaceX under the IPO structure. He will be CEO, CTO, and chairman of the board simultaneously after listing.

Any institutional investor buying SPCX shares is making a pure bet on one man's judgment — a man who is simultaneously running Tesla, xAI, The Boring Company, and who was until recently running DOGE inside the federal government. There are 36 pages of risk factors in this filing. Musk's centrality is one of the biggest.

The 1 Million Satellite Plan — and the $2 Trillion Problem

Musk posted a mission statement on SpaceX's website promising to launch one million AI data center satellites into orbit starting in 2028, using Starship. According to Forbes, the plan involves launching Starship once every hour to make it happen.

Robert Zubrin — an aerospace engineer and founder of the Mars Society, best known for the Mars Direct mission architecture — ran the numbers. Speaking to Forbes contributor Kevin Holden Platt, Zubrin said: SpaceX spent roughly $2 million per satellite building and launching its 10,000 Starlink broadband satellites. Apply that cost structure to one million AI satellites, and you're looking at $2 trillion.

That's the entire projected IPO valuation. Gone. To build the constellation.

Zubrin told Forbes this plan could "spark the lightning-speed halt" of Musk's winning streak. He also pointed out that orbital AI data centers would be far more expensive than terrestrial ones — meaning the economics don't work even if the rockets do.

None of the major financial coverage — not Bloomberg, not CNBC, not the Wall Street Journal — is centering this critique. They're running the valuation numbers and moving on.

What the Left-Leaning Coverage Gets Right — and Gets Wrong

NPR and TechCrunch do solid work pulling out the key financial figures.

But both outlets spend considerable energy on the "Musk gets richer" frame — NPR literally led with the line about him potentially becoming "the world's first trillionaire." That's not inaccurate, but it's also not the central question in this filing.

The central question is whether the underlying business can justify a $1.75 trillion valuation when its AI division — the bet that consumed 60% of capex last year — is massively underperforming. Neither outlet asked that question directly.

What Regular Investors Need to Understand

Starlink is real. Launch services are real. Those businesses generate cash.

But SpaceX is going to market as an AI and multiplanetary infrastructure company, not a satellite internet provider. That framing justifies the valuation. And that framing depends entirely on Starship working at industrial scale, AI satellites being economically viable, and Musk executing flawlessly across four or five simultaneous companies.

If any of those assumptions breaks — and Zubrin is telling you at least one of them is financially delusional — the valuation collapses.

The IPO is targeting mid-June, per NPR. It could raise up to $80 billion, which would make it the largest IPO in history by a significant margin — Saudi Aramco raised $29 billion in 2019.

The retail investor who buys SPCX on day one is betting that Musk is right about everything. That has sometimes worked out. It has also sometimes produced billion-dollar wipeouts.

Review the 36 pages of risk factors before investing.