SpaceX will host bankers, analysts, and institutional investors at an in-person analyst day on Tuesday, April 21 — the first disclosure event ahead of a confidentially filed IPO that targets a $1.75 trillion valuation and up to $75 billion in proceeds. [1] The math the room has been circulating for two weeks does not round well. At $1.75 trillion against 2025 revenue of roughly $16 billion, SpaceX would list at a price-to-sales multiple of approximately 109. Against EBITDA near $8 billion for the same period, the price-to-EBITDA ratio lands around 220 on trailing numbers — or 109 on Reuters's aggressive double-the-2026-revenue assumption, which is what the Reuters math piece used to make the number more defensible. [2] Tesla trades at 12 times revenue and 79 times EBITDA. Palantir trades at 43 and 75. SpaceX is asking the market to underwrite multiples that even the market's reigning AI narrative stocks have not been asked to defend.

The paper's Saturday framing named Monday as the week SpaceX's math met the ceasefire; the handover's calendar was off, and the event has landed on Tuesday. The frame holds. Oil re-opened the week above $96 Brent on Monday, Tesla's inventory overhang sits at 50,363 units heading into Wednesday's earnings, and Friday's Netflix close erased $44 billion in market capitalization on the guidance the market demanded. The macro tape is one where very large prices for growth narratives are being tested in real time. The SpaceX analyst day is an institutional disclosure scheduled to ask bankers to sign the book on what would be the largest IPO on record, and to do it two days after the market chose to punish a durable subscription business for an admission of growth ceiling.

Starlink is the only part of the math that rounds. Analyst consensus — Reuters, PitchBook, Bloomberg, TradingKey — puts Starlink's 2025 revenue at $10.6 billion to $11.4 billion and its EBITDA margin at roughly 54 to 63 percent. [3] That makes Starlink, on its own, 50 to 80 percent of consolidated SpaceX revenue and the entirety of consolidated profitability. At subscriber counts of 9.2 to 10 million by end-2025 and projected growth to 15.9 to 24 billion in 2026 revenue, the valuation room has a real business to defend. The question is whether Starlink, alone, can carry a $1.75 trillion price tag that Tesla's full enterprise — automobiles, energy, storage, robotaxi narrative, Dojo, Optimus — cannot carry at its current $450 billion equivalent. Elon Musk leads both companies; the banker math says one of them is mispriced against the other.

The Futurum analyst Shay Boloor, whose line that "Starlink is the only reason the valuation is defensible" circulated widely on valuation X this month, has been the most-quoted skeptic voice in Reuters's coverage of the filing. [2] His framework — widely echoed in institutional notes — is that the sum-of-the-parts case for $1.75 trillion requires assigning positive enterprise value to four businesses that have not yet proven profitable: Starship launch, satellite launch for third parties, the Starshield defense tier, and the xAI "Macrohard" Memphis datacenter that SpaceX absorbed in February's structural cleanup. The bear case is that three of those four are cost centers and the fourth is a money-losing AI operation consolidated via acquisition. Boloor is not alone; PitchBook's Franco Granda has been on the record in Reuters framing the valuation as asking the public market for credit on narratives the private market has already priced into the $1.25 trillion secondary round in February. [1]

The analyst day is structured in a sequence that itself tells a story. Tuesday's in-person session is the standard disclosure event. Thursday, April 23 brings a tour of the Memphis datacenter — the xAI "Macrohard" asset, the physical installation of the AI narrative the company acquired to justify valuation multiples traditional aerospace could not produce. Two weeks later, on May 4, SpaceX will host a virtual session to walk research teams at sell-side banks through the financial models. [4] That sequence is the unusual part. An analyst day ordinarily includes the financial models. A company that is confident its models will hold up in the room shows them in the room. SpaceX's chosen sequence — show the story in person, show the hardware on a field trip, walk through the numbers virtually two weeks later — implies the models need more time than the in-person session provides, or more flexibility than the sell-side would permit if the walkthrough happened Tuesday.

The second structural feature of this week's disclosure is that SpaceX has not published 2025 audited financials. Revenue and EBITDA are known from Reuters's January scoop ($15-16 billion; $8 billion EBITDA) and CFO Bret Johnsen's presentations to IPO bankers (Starlink's addressable market at $1.6 trillion, total space business at $370 billion). [5] The confidential SEC filing happened on April 1. The prospectus is expected in April or early May. The 15-day wait period follows, then the roadshow. A June listing remains the target. Institutional buyers will have seen the models at the May 4 virtual session, will have written the work by mid-May, and will be expected to commit book by the roadshow. The window is compressed.

The market the book is walking into is priced for peace. Friday closed oil down roughly 9 percent on the week — the largest weekly decline of the war — before Sunday's Touska seizure and Iran's strait re-closure partially reversed the pricing. The six major American banks has already reserved against war; the ceasefire-rally tape has already paid some of those reserves back. Asking a book to commit to a $1.75 trillion cut of that equity right now is not asking the market to extend itself during a bull run. It is asking during a week when the market has been given two conflicting signals — ceasefire priced into Friday, kinetic priced into Sunday — and must choose which one the SpaceX valuation is walking into.

The paper's Saturday framing was that the Tuesday analyst day would be the first institutional test of whether the peace-priced market can underwrite the biggest IPO in history. The Sunday kinetic has complicated the framing without resolving it. If the book commits Tuesday at 80x sales, the peace trade dominates; if bankers defer to the May 4 virtual session, the market is buying time to see whether Starlink's subscriber growth trajectory and the xAI datacenter's narrative can carry the gap the room cannot close in person. Reuters's April 1 reporting that the IPO would be the largest on record is operative only if the book commits. The failure mode — publicly — is that the book is softer than the filing, the roadshow is moved out, the price is cut. Elon Musk has taken companies public before. He has never taken one public at 109x revenue on a Tuesday two days after the market discovered his valuation's closest peer could lose $44 billion in one session.

Four trading days before the lid comes off, the paper is watching for the artifact that will signal which way Tuesday goes: whether Musk himself shows up to the in-person session or appears virtually; whether CFO Johnsen takes questions on the model; whether the May 4 virtual walkthrough is accelerated or canceled. Each of those would move the signal materially. For Monday's open the market has one piece of operational information: the analyst day is Tuesday, not Monday, and the financial models are fourteen days out.

-- THEO KAPLAN, San Francisco