I bought the IPO for 2030.
Then it signed a contract seven days before the opening bell that never made it into the prospectus.
I bought this IPO in June. I’ve watched every Starship flight live, all thirteen, and I won’t pretend the rockets had nothing to do with it. After all I want to participate in the Star Trek future that my childhood showed me.
But the filed IPO document was cold. No revenue guidance. None.
Search the S-1, both amendments, and the final prospectus and “annualized recurring revenue” returns zero hits. No $100 billion figure. No forward projection of any kind.
You were buying $18.674 billion of 2025 revenue, a $2.6 billion operating loss, and a TAM slide.
I bought 2030 and I was fine being early.
Then the company reported its first public quarter, and I came off that call more bullish than I went in. Not about Mars - about the next five months.
Here’s the part that jumped out.
The prospectus disclosed one compute contract. Anthropic, signed May 3 — about 325,000 Nvidia GPUs at $1.25 billion a month through May 2029.
That was in the book and priced. Google was not.
SpaceX signed Google on June 5. After the original S-1. After both amendments. Seven days before it started trading.
It was disclosed in a one-page Rule 433 filing and never made it into the final June 12 prospectus at all. The compute section in the 424B4 still describes only Anthropic.
Roughly 110,000 GPUs. $920 million a month, October 2026 through June 2029.
Then in the first three weeks of July, after the quarter had already closed, the CFO says they signed another $6.7 billion of cloud services revenue.
Stack them together. Google plus Anthropic is $2.17 billion a month. About $26 billion annualized — and Google’s hasn’t started yet.
Add the July signings ramping from October. Add Starlink: 12 million subscribers, $66 blended ARPU, the best net-add quarter in company history.
That’s roughly $75 billion of visible annualized revenue by December, and nearly all of it is already signed.
Musk says December is $100 billion. Okay - his track record isn’t fantastic on timing, but with what we’ve seen thus far it seems plausible.
Of course, the market disagreed with the importance of this information when it dumped the stock 13.6% to below its $135 IPO price.
The full breakdown is on the site — the $18.4 billion of quarterly capex run as a payback calculation instead of a burn number, the per-watt monetization figure Musk put on the record that makes the last $25 billion a contract question rather than a capacity question, what I own and where I’m adding, and the one clause buried in both compute contracts that would turn the whole setup from a purchase order into a hole.
The market spent post-earnings Wednesday marking down a company for buying the thing it sells, on a sub-twelve-month payback, with the customers already signed.
I bought the rockets and the future. The data centers are what pay the SPCX shareholders today.
Thank you for reading.
Read the whole thing 👉
https://evervests.com/articles/spacex-may-not-need-mars-to-justify-the-valuation
🔗 Join the growing list of informed investors at Evervests.com
Not financial advice. Full disclosures on evervests.com
.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.