The $1.5 Trillion Rocket: Inside the SpaceX IPO Valuation

Your premise, fact-checked

You asked how a company that makes zero profit can be worth half the aerospace sector while trading at thousands of times earnings and delivering "next to nothing outside Starlink." Four claims — here's how each holds up against SpaceX's S-1 filing.

The profit picture — and the xAI twist

SpaceX's 2025 income statement looks contradictory until you see what's inside it. Revenue grew 33% to $18.7 billion. Adjusted EBITDA was a healthy +$6.6 billion. Starlink alone threw off $4.4 billion in operating profit. And yet the company reported a $4.94 billion net loss — followed by a $4.28 billion loss in Q1 2026 alone.^[1][7]

The reconciliation is one acquisition. In February 2026, SpaceX merged with xAI — Elon Musk's AI company, which had itself absorbed X (formerly Twitter) in 2025 — in a ~$1.25 trillion deal, the largest merger ever.^[5][12] xAI lost $6.4 billion in 2025 on just $3.2 billion of revenue, and SpaceX poured $12.7 billion of capex into AI data centers (the "Colossus" complex in Memphis).^[4] A company that earned a $791 million profit in 2024 became one that loses billions — by choice.

So the honest framing of your "zero profit" point: the space business is profitable; the consolidated company is not, because xAI's losses exceed everything Starlink earns. The other GAAP drags — heavy stock-based compensation and depreciation on the Starlink constellation — are real too, but xAI is the swing factor.^[4]

Half the sector: the valuation in context

This is the claim that sounds like hyperbole and isn't. The entire publicly-traded US aerospace-and-defense sector — Boeing, RTX, Lockheed, all 373 listed names — is worth about $2.89 trillion.^[11] A single, still-private company is about to list at roughly half of that.

Put differently: you could buy RTX, Boeing, Lockheed, General Dynamics, Northrop, and L3Harris outright — every flagship of American defense — and still have less than half of SpaceX's price tag.^[9] Bloomberg-style framing aside, that is a genuinely unprecedented concentration of value in one unproven public equity.

The multiple: ~80× sales, ~1,900× profit

Valuation multiples translate the price into "how many years of … am I paying for." For SpaceX the answers are extreme.

At $1.5 trillion on $18.7 billion of revenue, SpaceX is priced at about 80× sales — and 95× at the $1.77T figure in the S-1.^[2] Traditional primes trade around 2–3× sales. On earnings the comparison breaks entirely: a GAAP net loss means there is no price-to-earnings ratio. Measured against the company's last positive annual profit — the $791 million it earned in 2024 before xAI — the price is roughly 1,900 times earnings. For reference, the dot-com peak that later cratered saw multiples a fraction of that on far smaller companies. This is a growth-and-optionality bet, not an earnings bet.

The deadweight: SpaceX × xAI × Twitter

You called it "the deadweight of Twitter," and the corporate plumbing proves the instinct. The IPO entity is a Russian doll of Musk companies:

• SpaceX — rockets (Falcon 9, Starship) and Starlink (satellite internet). Profitable.
• xAI — the "Grok" AI models. Burning ~$6.4B/year. Merged into SpaceX Feb 2026.^[5]
• X (Twitter) — acquired by xAI in 2025, so it rode into SpaceX with the merger. The S-1 is literally "the first public glimpse into X's financials."^[4]

By May 2026, Musk said xAI would cease to exist as a standalone company, folding Grok and X into a SpaceX AI unit.^[12] The stated logic is infrastructure convergence — Starlink's satellites plus orbital data centers to dodge the terrestrial power limits throttling AI.^[12] The unstated logic is financial packaging: bundle a profitable, beloved rocket company with a loss-making AI venture so investors value them as one story. When you buy SPCX, you are buying Twitter's losses whether you want them or not.

What it actually provides

Strip away the AI and the question becomes: is the space business alone worth a fortune? Here's what it really delivers.

Launch: a near-monopoly that mostly serves itself

In 2025 SpaceX flew 165 Falcon 9 missions — more than China (90), Russia (17), France (7), India (4) and Japan (3) combined.^[14] Reusability gives it a structural cost edge: Falcon 9 lists at ~$74M versus $100M+ for Arianespace or ULA.^[14] The catch you identified is real, though — only 43 of those 165 launches flew for outside customers.^[3] The other ~74% were launching Starlink's own satellites. The launch business is less a revenue engine than the moat that makes Starlink cheap to build.

Starlink: the actual cash machine

Starlink booked $11.4 billion in revenue (61% of the total), grew ~50% year-over-year, and runs a 63% EBITDA margin.^[3] Subscribers went from 2.3M (2023) → 4.4M (2024) → 8.9M (2025), passing 10 million by February 2026.^[3] One caveat the bears flag: average revenue per user fell from ~$99 to ~$66 as growth shifted to cheaper markets — volume is rising, price is falling.^[2]

Starship: optionality, not yet income

The launch/Starship segment has the best gross margins (~67%) but grew revenue only ~7.6% in 2025.^[2] Its value in the IPO is almost entirely future-tense: heavy lift, Mars, and the orbital infrastructure the AI story depends on.

The bull case vs. the bear case

The disclosure made the story bigger, but also more volatile… my concern is that SpaceX will overreach in the AI business. — Aswath Damodaran, NYU Stern (the "Dean of Valuation")^[6]

Verdict: is it worth it?

Holding the line on a clear call rather than a shrug: at $1.5–1.77 trillion the price is hard to defend on anything but faith. The strongest steel-man for the valuation is that SpaceX owns an irreplaceable launch monopoly and the fastest-scaling satellite ISP on Earth — both genuinely rare, genuinely defensible franchises. But even crediting those generously, the most rigorous independent valuation on the public record (Damodaran's ~$1.25–1.3T) lands below the IPO range, and that estimate still has to swallow a $6-billion-a-year AI loss whose payoff rests on a market size its author calls fantasy.^[6]

So your underlying suspicion is sound, with the nuance that makes it sharper: this isn't a profitless company being valued at half a sector on nothing. It's a profitable space company being valued at half a sector — and then having a profitless AI company stapled to it, so the combined entity reports losses while the price assumes Mars, global connectivity, and a winning AI lab all pay off at once. You're not overpaying for rockets. You're paying rocket-monopoly prices and getting Twitter's losses, Grok's bet, and Musk's control in the same envelope.

The case against this verdict: launch and Starlink are the kind of winner-take-most franchises that have humbled valuation skeptics before — if Starship and direct-to-cell Starlink compound the way Falcon 9 did, today's "absurd" multiple could look cheap in a decade. That's the bet. Just know that it is a bet, priced near perfection, with no dividend, a lock-up, and one man holding the wheel.