SpaceX Reveals Plans for Orbital Data Centers in IPO Prospectus
In its IPO prospectus, SpaceX announced plans to deploy orbital solar-powered data centers to support AI. The company also disclosed an existing contract with Anthropic that will generate $45 billion in revenue by 2029, and the IPO itself could make Elon Musk the world's first trillionaire.
Trillionaire on a Space Leash: Why $45 Billion from Anthropic Isn't the Main Story in SpaceX's Prospectus
Opinion by an independent analyst, May 25, 2026
On May 20, SpaceX finally unveiled its IPO prospectus. The media immediately latched onto three big hooks: orbital data centers, a $45 billion contract with Anthropic, and the prospect of making Elon Musk the first trillionaire in history.
But as an analyst who has watched Musk's deal structures for years and knows how to read between the lines of a 400-page S-1, I'm here to tell you: you're looking in the wrong place. The main story isn't about money from a competitor. And it's not about space.
The main story is about absolute, legally unassailable shareholder dictatorship that Musk just patented for the entire public company market. And about how Starlink is actually drowning in its own pricing, while the Anthropic revenue is not a victory but an admission of Grok's failure.
[The Gist]: Who Really Pays and Who Gets Paid
What happened? SpaceX filed for an IPO under the ticker SPCX on Nasdaq, with a target listing date of June 12 and a valuation of $1.75-2 trillion.
The document outlines three business segments:
- Starlink (communications): $11.4 billion revenue in 2025, $4.4 billion operating profit.
- Space (launches): $4.1 billion revenue, $660 million loss.
- xAI/X (AI): $3.2 billion revenue, $6.4 billion loss.
Anthropic, a competitor to xAI, is leasing SpaceX data center capacity (Colossus I and II in Memphis) for $1.25 billion per month until May 2029. That totals about $45 billion in revenue.
But here's what they're not telling you: either party can terminate the contract with 90 days' notice. This isn't a "guaranteed cash flow"—it's an option. Anthropic is paying that much because it has no choice: capacity is under Musk's control, and AWS and Google can no longer meet demand.
Non-obvious insight:
The Anthropic contract isn't monetization of SpaceX's AI infrastructure. It's an official admission that Musk's own chatbot Grok is a commercial corpse. If Grok generated demand, Musk wouldn't be leasing capacity to a competitor, especially one funded by Amazon and Google. xAI's annual loss is $6.4 billion. Grok doesn't make money. Instead, SpaceX has become a hosting provider for Anthropic. Musk is essentially renting out his house to a neighbor because he can't pay the utility bills himself.
Timeline and Context: How Musk Set the Stage
- February 2026: Musk announces the merger of SpaceX, xAI, and X (Twitter) into a single entity valued at $1.25 trillion. This was a preparatory step for the IPO—needed to consolidate losses so that later he could "take out the trash" and show investors only Starlink.
- May 2026: Ten days before filing the S-1, SpaceX files an application with the FCC to launch up to 1 million satellites for orbital data centers.
- May 20, 2026: Publication of the prospectus. In it, Musk secures 93.6% of voting Class B shares (10 votes per share).
- June 4-12: Roadshow, pricing, listing.
Key date: June 12, 2026. On that day, a company will appear on Nasdaq where firing the CEO (Musk) is technically impossible. The board of directors is a formality. Public shareholders will get voting rights, but their stake will be diluted to a negligible level.
Who Wins and Who Loses
Big winners:
- Elon Musk. His 41% stake at a $1.75 trillion valuation is worth $717 billion. Plus Tesla, xAI, X, Neuralink—he becomes the first trillionaire in human history. But more importantly, he just set a precedent. Now any founder of a public company can cite the "Musk precedent" and ask the board for a similar control structure.
- Early investors: Peter Thiel (Founders Fund), Antonio Gracias (Valor Equity Partners), Luke Nosek. Their stakes are worth billions—provided they wait out the lock-up period (366 days for Musk, 180 for some others).
- Goldman Sachs, Morgan Stanley, Bank of America: The underwriters will earn hundreds of millions in fees.
Big losers:
- Public (retail) shareholders. They will buy shares at $??? (pricing on June 11) but will get neither control nor influence. They are passive capital providers under absolute dictatorship. If Musk decides to spend $100 billion on Mars, they can't do anything. If he decides to burn another $50 billion on Grok 3.0—same.
- Anthropic. Yes, they got capacity. But they just made their main supplier a direct competitor. And any whim of Musk could leave Claude without GPUs in 90 days.
What the Media Isn't Saying
Three things you won't read in the headlines:
1. Starlink is in a price war it's losing.
Average revenue per user (ARPU) for Starlink has dropped from $99 in 2023 to $66 in Q1 2026. That's a 33% decline in three years. Officially, it's a "strategic expansion into developing countries." In reality, competition from Amazon Kuiper and Chinese constellations is forcing Musk to undercut prices. Subscribers have grown to 10.3 million—yes, but each new subscriber pays less than the previous one. The model scales only downward in margin.
2. Orbital data centers are technologically infeasible in the next 5 years.
The prospectus has nice words about "vacuum cooling of space and solar energy." But calculations by engineers at Varda Space Industries (cited in the S-1 but buried in fine print) show that orbital computing currently costs three times as much as terrestrial. Even with Starship at $10 million per launch, the payback period for an orbital data center is 12-15 years. No commercial client will pay triple for the same thing just because "it's in space." This entire section of the prospectus is marketing for valuation.
3. Mars Colony is not a business plan—it's a condition for Musk's bonus.
The prospectus stipulates: Musk will receive another tranche of shares (67 million) only if three conditions are met: (a) a Mars colony with a population of 1 million, (b) orbital data centers with 100 terawatts of capacity, (c) a market cap of $7.5 trillion. This isn't a plan. It's a science fiction novel legally embedded in the document to justify 93.6% control. None of these triggers will be achieved in Musk's lifetime (except maybe the market cap, but $7.5 trillion is more than Apple, Microsoft, and Nvidia combined).
Forecast: Next 30 and 90 Days
30 days (June 2026):
June 11—IPO pricing. Expected range: $55-65 per share (valuation $1.6-1.8 trillion). June 12—first trading day. Expect wild volatility: opening up 15-20%, then a 5-7% correction within 48 hours. Institutions will get the bulk of allocations; retail will get crumbs. My recommendation: don't participate in the first week of trading. Better to wait for a pullback after lock-up expiry (180 days), when early investors start taking profits.
90 days (August 2026):
Key metric to watch is not the SPCX price, but Starlink ARPU and subscription growth in India and Africa. If ARPU falls below $60, shares will correct 20-25% from the IPO price. Second factor: will Anthropic terminate the contract? (90 days from May means they could give notice in August.) If they do, it would signal that the AI boom is slowing, and SpaceX would lose $1.25 billion in monthly revenue that hasn't even materialized yet.
Editorial Forecast
Asset: SpaceX shares (SPCX) — first 72 hours after listing on June 12.
Direction: Surge in the first hours (up to 15-20% from the offering price), then a pullback of 5-8% by the end of the second day.
Key levels: Offering price approximately $60. Resistance at $68-70, support at $54-55.
Confidence: Medium (60%) — IPOs of this scale are unpredictable, especially with a 5% free float.
Main risk: A technical glitch on Nasdaq on listing day (non-zero probability given the system load) or a sudden SEC announcement suspending trading due to "corporate governance concerns." If a major institution (e.g., Fidelity) publicly states it won't participate due to Musk's concentration of power, shares could crash 10% in the first hour.
This is an editorial opinion and not investment advice.
— Editorial Team
