SpaceX Q3 Earnings: $14.6 Billion in Revenue Beat the $12.7 Billion Estimate. Here's What the Numbers Prove, What They Don't, and What Musk Said About Tesla.

Honest fence (read first): SpaceX (Nasdaq: SPCX, public since its June 12, 2026 IPO) released its Q3 2026 results after the market closed on Tuesday, Oct 20, and held its webcast at 4:30 PM ET. Company numbers come from its SEC filing (8-K, Exhibit 99.1). Analyst estimates come from LSEG (via Reuters and CNBC), Bloomberg (via Yahoo Finance) and StreetAccount (via CNBC). Call quotes come from Benzinga's automated transcript and were cross-checked against CNBC's live coverage. The official replay and transcript are on ir.spacex.com. Prices were last checked after the close on Thu Oct 22 and will have moved by the time you read this.

Editorial Thesis

The headline number is real, and it's big: $14.6 billion in revenue against a about $12.7 billion estimate. The filing backs that up. What the filing can't back up yet is the story told around it: 13 gigawatts of compute next year, $100 billion to $1 trillion in annual recurring revenue by the end of 2027, and weekly Starship flights. A learner's job is to keep those two piles separate.

Honest Verdict

Proven (filed): revenue, profit, Starlink subscribers, compute capacity, capex and the balance sheet. Promised (said on the call): the Q1 2027 revenue range, Q4 capex, the compute ramp, Starship cadence, and every ARR number. Unproven: whether the AI spending keeps paying off at today's rental prices, and whether a Tesla merger happens at all. Musk himself said it's not a certainty.

The headline numbers (all from SpaceX's 8-K unless noted)

• Revenue: $14,575 million, up 161% from $5,576 million a year earlier and up from $7,814 million in Q2. LSEG estimate: $12.72 billion (Reuters/CNBC). Bloomberg consensus: $12.7 billion (Yahoo Finance). That's about 15% above the LSEG number (our math).
• Net income: $4,404 million, compared with a $1,054 million loss a year ago. LSEG expected about $1.57 billion (Reuters).
• Earnings per share: $0.35 diluted (GAAP) in the filing. LSEG/IBES counts 45 cents adjusted vs a 13-cent estimate (CNBC, Reuters). Bloomberg's comparison was $0.35 vs $0.15 (Yahoo). The two "beat" sizes differ because they measure different things.
• Operating income: $2,791 million, compared with a $273 million loss a year ago.
• Adjusted EBITDA: $8,809 million (non-GAAP), up 306%. Bloomberg consensus was $6.59 billion (Yahoo).
• Capex: $19,593 million in the quarter, roughly in line with the about $19.6 billion consensus (Yahoo). AI capex was $16.44 billion vs $16.9 billion expected.

Where the money came from

• Super Intelligence (SI, formerly the AI segment): $7,560 million, up from $813 million a year ago (+830%) and $2,561 million in Q2. StreetAccount expected $6.47 billion (CNBC). Most of it is "SI solutions & infrastructure" ($7,022M): renting compute to customers such as Anthropic, Google and Reflection AI (Reuters), plus Cursor and Grok. Cursor added $608 million after the deal closed on Aug 14.
• Connectivity (mostly Starlink): $5,248 million, up 69%. StreetAccount expected $4.84 billion (CNBC). Consumer revenue was $3,121M and enterprise & government $2,127M (+91%).
• Space (rockets): $1,767 million, up 6% from a year ago and 84% from Q2.

The positives

1. A clean revenue beat, filed with the SEC. $14.6B against about $12.7B, with all three segments ahead of the StreetAccount segment estimates CNBC reported.
2. Profit at every level of the income statement. It was the first profitable quarter as a public company (WSJ). The SI segment swung to $431 million of operating income from a $1,257 million loss in Q2. The CFO says that's about $1.6 billion before $1.2 billion of Cursor-related stock compensation.
3. Starlink beat on subscribers. 13.6 million consumer subscribers, up 1.6 million in the quarter. Analysts expected 12.19 million (Reuters). Consumer ARPU rose to $69 a month, from $66 in Q2. Connectivity operating income nearly doubled, to $2,367 million. More than 3,800 aircraft had Starlink as of Sep 30.
4. Compute and backlog. Nameplate compute reached 1.7 GW (0.6 GW a year ago). The CFO said on the call that backlog was $63 billion at Sep 30, with about $42 billion expected to be recognized within a year.
5. Big guidance (a promise, not a result). CFO Bret Johnsen guided Q1 2027 revenue to $33 billion to $37 billion. LSEG's average estimate was $22.2 billion (CNBC). He also repeated the goal of $100 billion in ARR by the end of 2026, based on December revenue annualized.
6. Starship progress (filed fact plus a plan). Flight 14 was Starship's first orbital flight, delivering 26 Starlink V3 satellites. SpaceX says that's about 10 times the downlink capacity of a Falcon 9 Starlink launch. On the call, Musk said Flight 15, which he expects next month, will try to catch both the booster and the ship, and that success is "not guaranteed."
7. Liquidity. Cash of $51.3 billion plus $38.6 billion in marketable securities, about $89.9 billion in total (our addition).

Market reaction (not a forecast): SPCX closed at $147.79 on Oct 20 and traded around $154.50 after hours (+4.5%) at 5:35 PM ET (StockAnalysis). CNBC reported a gain of about 5% in extended trading. After-hours moves often reverse.

Price update (Thu Oct 22, after the close): SPCX closed Wed Oct 21 at $159.31, up about 7.8% from $147.79. On Thu Oct 22, the day the first post-earnings lock-up batch (up to 1.3 billion shares) became eligible to sell, it fell about 9.1% to close at $144.75, with a day range of $143.21 to $154.22 (Yahoo Finance). That gave back the whole post-earnings jump. One day's move doesn't predict the next.

The negatives and risks

1. Spending is far bigger than cash coming in. Capex was $48.1 billion in the first nine months, against $13.1 billion of operating cash flow. That's about $35 billion short (our math, from the cash-flow statement). The CFO guided Q4 capex to $35 billion to $40 billion, roughly double Q3.
2. More debt. Debt and finance leases rose to $38.8 billion from $22.9 billion at Dec 31, 2025 (our addition of current and long-term). About $12.9 billion of that is labeled related-party, and the release doesn't name the party. Morgan Stanley's Adam Jonas called financing "the biggest risk" to his model before the report (Benzinga, Oct 20). On the call, the CFO pointed to operating cash flow, investment-grade debt markets and vendor and customer partnerships.
3. The net income number has help. About $1.6 billion of the $4.4 billion came from below the operating line: $1,025M of other income ($561M of it related-party), $873M of interest income net of $675M interest expense, and a $390M tax benefit (our math from the filing). Operating income was $2.8 billion.
4. Margins should cool. The CFO said Q4 gross margins should look more like Q2's, because Q3 had one-time benefits and a favorable customer mix. Connectivity's 55% gross margin included about 3 points from tariff refunds.
5. Launch cadence fell. 35 launches in Q3 vs 44 a year ago. Year to date it's 113 vs 128. Mass to orbit was 476 tonnes vs 525. The Space segment posted a $(7) million operating loss (vs +$259M a year ago) as Starship R&D rose to $1,225M.
6. Starlink ARPU is lower than a year ago. It's $69 now vs $83 in Q3 2025, as SpaceX grows abroad with cheaper plans (Reuters).
7. AI revenue depends on renting compute. Most SI revenue comes from cloud contracts. Musk said SpaceX rents "at effectively the spot market rate," not long-term deals. That helps if prices rise and hurts if they fall. CNBC noted neocloud peers CoreWeave and Nebius are down about 28% in October.
8. Stock-comp overhang from Cursor. SpaceX says $23.4 billion of Cursor-related equity awards will be expensed over roughly three years. Q3 alone took $1,217 million.
9. Big numbers on the call are claims. Musk said ARR by the end of 2027 is "somewhere between 100 billion and a trillion," and that he's "fairly confident" of adding 10 GW of compute in 2027. Neither is in the filing. Treat them as goals.
10. Other headline risk. CNBC cites a WSJ report that Musk ordered Starshield restrictions in Ukraine. That's a reported story, not a filing.

Lock-up math: Thursday Oct 22, then Saturday Oct 24

Update: the first batch, up to 1.3 billion shares, became eligible to sell on Thursday, Oct 22. The stock fell about 9% that day (see the price update above).

From the 424B4 prospectus (verified against the filing text):

• "The second full trading day on Nasdaq immediately following the public release of our quarterly financial results for the quarter ended September 30, 2026: up to 1.3 billion shares of Class A common stock, representing 28% of the shares subject to the 180-day lock-up period." Results came out after the close on Tue Oct 20. Wed Oct 21 is the first full trading day after that, and Thu Oct 22 is the second. That's our reading of the prospectus wording.
• "October 24, 2026 (135th day): up to 328.4 million shares... 7%." Oct 24 is a Saturday, so Mon Oct 26 is the first session those shares can trade.
• "Up to" means eligible, not sold. Some holders sell and many don't. Affiliates face extra limits. Musk's own shares stay locked until June 2027. In August, CNBC notes, the stock rose after the first unlock rather than falling. Past moves don't predict the next one.

Tesla merger: what was actually said

It came up on the call. Deutsche Bank's Edison Yu asked about the pros, cons and roadblocks of a combination. Musk's answer, per the Benzinga transcript and CNBC:

"Both SpaceX and Tesla have formed special committees to discuss this idea at the request of shareholders. These special committees consist of independent directors, and because I'm a conflicted party, I can't actually be in the discussions of the special committees... some people out there assume that a SpaceX and Tesla merger is a certainty, and I think that would be overstating it. Do I think it would be a good idea? Yes. But at the end of the day the shareholders have to be supportive of this, and particularly on the Tesla side we'll need a 50% vote of all shares , not just those who vote, but of all shareholders... at this point the special committees have not reached agreement on the terms of any transaction, and obviously there's no guarantee that they will."

What that means: committees exist and are talking. There are no agreed terms, no announced deal, and no guarantee of one. A Tesla vote would need a majority of all shares. Outside views are only opinions: Futurum's Daniel Newman said a merger "feel[s] imminent," and Barclays suggested a deal could be announced before Tesla's annual meeting (Reuters). No deal exists today. On Tesla's own Q3 call on Wednesday, Oct 21, Musk added nothing new: "I think I said everything that should be said yesterday" (CNBC, Motley Fool transcript).

Reaction on X

• @SpaceX posted the release and a highlights thread ("Delivered revenue of $14.6 billion, up 161% year-over-year").
• Sawyer Merritt and others quickly quoted the merger answer. The Kobeissi Letter led with the committees.
• @elonmusk: as of about 5:50 PM ET we found no standalone Musk post about the numbers. He replied with a laughing emoji to a shareholder's joke that "SpaceX shareholders had to bail out xAI." That's a reaction, not information.

Claim vs proof

Video note

Rob Maurer (Tesla Daily), "Live: SpaceX Q3 Earnings Call 2026 (SPCX)" (xetd5bxGPe4, 1 hr 41 min, posted Oct 20). It carries the full earnings call with Elon Musk, COO Gwynne Shotwell and CFO Bret Johnsen, followed by Maurer's own review. Disclosure: Rob Maurer says in the video description that he owns SPCX stock. His views are his, not ours.