SpaceX’s first public print has no precedent — but the S-1 already graded it: $18.67B revenue, an AI segment losing $6.4B, a $19.6B spectrum bet
Tuesday, August 4, SpaceX (SPCX) prints its first quarter as a public company (~4:30 p.m. ET; call ~5:00 p.m. ET) — and here is the problem with grading a maiden print: there is no earnings history to anchor on, only the S-1. Everyone will react to whatever number tops the release; the S-1 already tells you which numbers matter and why the AI segment even exists. Graham Corporation (GHM) already filed full fiscal 2026 results on June 8 — vacuum, heat-transfer, and fluid systems into Defense, Space, and Energy. Read them as a stack: platform company on the call, supplier backlog already on EDGAR.
This is research, not investment advice — no price calls here. Figures from SEC S-1 / Exhibit 99.1 filings.
Who prints when
| Ticker | Status | Last filed checkpoint |
|---|---|---|
| SPCX | Aug 4 ~4:30 p.m. ET (maiden public) | S-1 / prospectus financials (FY25 + Q1'26) |
| GHM | Already filed Jun 8 | FY2026 (ended Mar 31) Ex 99.1 |
SpaceX — S-1 baselines into the first public print
From the IPO disclosure package (CIK
0001181412;
S-1 accession 0001628280-26-036936; prospectus family including
0001628280-26-040364):
Consolidated
| Line | FY2025 | Q1 2026 |
|---|---|---|
| Revenue | $18.674B | $4.694B |
| Loss from operations | $(2.589)B | $(1.943)B |
| Adjusted EBITDA | $6.584B | $1.127B |
Segments (the real scorecard)
| Segment | FY2025 revenue | FY2025 Seg. Adj. EBITDA | Q1'26 revenue |
|---|---|---|---|
| Connectivity (Starlink) | $11.387B (+49.8% YoY) | $7.168B (+86.2%) | $3.257B |
| Space (launch & related) | $4.086B | $653M | $619M |
| AI | $3.201B | $(1.237)B | $818M |
Connectivity also printed FY25 operating income of $4.423B (+120% YoY). AI printed FY25 operating loss of $(6.355)B — that is the dilution story the public market gets to re-price every quarter now.
Two S-1 disclosures that frame the whole print
Both from the S-1 subsequent-events disclosure (Item 15, accession
0001628280-26-036936):
- Why there is an AI segment at all: the xAI merger was consummated on February 2, 2026. The $(6.355)B FY25 AI operating loss is the acquired business the public company now carries — every quarter is a referendum on whether that loss curve bends.
- The spectrum bet: the pending EchoStar spectrum transaction totals approximately $19.6B — roughly $11.1B in equity (52.4M Class A shares at $212/share) plus up to $8.5B of EchoStar debt payoff, expected to close around November 2027. That is committed capital on the balance sheet's future, disclosed before the first earnings call.
Aug 4 scorecard: (1) Starlink revenue / ARPU / subscriber trajectory vs the S-1 run-rate, (2) Space segment cadence and Starship CapEx color, (3) whether AI segment losses are still expanding post-xAI, (4) consolidated Adj. EBITDA bridge vs the $6.58B FY25 bar, (5) free-cash-flow / CapEx disclosure quality on a live call — with the $19.6B EchoStar commitment as the funding backdrop.
Graham — already filed (supplier checklist for the same tape)
FY2026 Exhibit 99.1 filed Jun 8, 2026 (accession 0001104659-26-058914, CIK
0000716314):
| Line | Amount |
|---|---|
| Q4 revenue | $67.1M (+13%) |
| FY revenue | $245.3M (+17%) |
| FY diluted EPS / adj. EPS | $1.12 / $1.40 (+13% adj.) |
| FY Adjusted EBITDA | $26.0M (+16%) |
| FY orders | $359.4M (record) |
| Book-to-bill (FY) | 1.5x |
| Backlog (YE) | $532.6M (+29%) |
| Backlog mix | Defense ~85% |
| Backlog conversion guide | ~35–40% in next 12 months |
CEO commentary flagged Defense strength and building Space momentum as drivers of the record quarter. FY2027 guide frame in the release: Adjusted EBITDA $35–40M.
Graham is not reporting Tuesday. It is the filed upstream checklist: if SpaceX talks constellation CapEx denser or defense launch work hotter, GHM’s $532.6M backlog — roughly 85% Defense — and 1.5x book-to-bill are the small-cap numbers that already showed demand arriving at the vacuum-and-thermal layer.
How the two connect
SpaceX is the vertically integrated launch + Starlink + AI-compute platform. Graham sells engineered systems into Defense / Space / Energy programs. Same thematic tape, different risk: SPCX is first-public-print volatility against S-1 history; GHM is multi-year backlog conversion (mostly Defense) already on file.
One scoreboard for Tuesday
- SPCX Connectivity — hold the Starlink growth engine near the S-1 pace.
- SPCX AI losses — stabilize or keep expanding, post the Feb 2 xAI close?
- SPCX Space / Starship — cadence and CapEx honesty on a live call, with ~$19.6B of EchoStar commitments in the background.
- Read GHM backlog as confirmed upstream demand — not as a substitute print.
Headlines will move the stock Tuesday night; the S-1 is what sizes the exposure. Live multiples: SPCX · GHM.
Primary sources
SpaceX S-1 (CIK 0001181412):
https://www.sec.gov/Archives/edgar/data/1181412/000162828026036936/
(0001628280-26-036936); prospectus family incl. 0001628280-26-040364.
Graham FY2026 Ex 99.1:
https://www.sec.gov/Archives/edgar/data/716314/000110465926058914/tm2617569d1_ex99-1.htm
(0001104659-26-058914).
First step: before the call, open the S-1 segment table and write down two numbers — Starlink’s $3.257B Q1 revenue and the AI segment’s FY25 $(6.355)B operating loss. Grade everything Tuesday against those two.
Research, not investment advice.