SpaceX was scheduled to report third-quarter results after the close of trading on Tuesday, and as of Tuesday afternoon in New York the company had published nothing. What the market had instead was a consensus: 12.72 billion dollars of revenue and 13 cents a share of earnings, according to LSEG, against 7.8 billion dollars and a net loss of 541 million dollars in the second quarter.
The revenue figure would be up about 63 percent from the second quarter, on top of the 92 percent year-over-year growth reported then. Adjusted EBITDA in the second quarter was 3.5 billion dollars. This is the second set of numbers SpaceX has produced as a public company, and the first with a full quarter of segment disclosure to model against.
The composition matters more than the total. StreetAccount’s estimates put AI revenue at 6.47 billion dollars, connectivity at 4.84 billion dollars and the launch and space business at 1.41 billion dollars. AI revenue would be up from 2.56 billion dollars in the second quarter, an increase of roughly 146 percent, and AI adjusted EBITDA is expected at 3.32 billion dollars against 1.15 billion dollars.
That is the line the quarter turns on. “AI is the largest swing factor for 3Q26 revenue and EBITDA,” Brett Linzey of Mizuho wrote in a note on Monday, with an outperform rating and a 200 dollar price target. The company closed its 60 billion dollar acquisition of the AI coding company Cursor during the quarter, which will contribute revenue that did not exist three months ago and will also bring software margins that look nothing like the margins on rented graphics processors.
The connectivity business is the counterweight. Starlink is where SpaceX has historically made most of its revenue and profit, and the growth rate there is slowing. Timothy Horan of Oppenheimer told clients on Tuesday that he expects an in-line quarter with guidance reiterated, and that industry subscriber growth is slowing with limited room for price increases.
The focus, he said, is shifting from net additions and service revenue toward cost cuts, lower subsidies and less churn. Saxo’s preview put consensus operating income for connectivity near 1.6 billion dollars.
Starlink’s position changed this month in ways that will not appear in the third-quarter numbers: SpaceX agreed to buy 800 megahertz of low-band spectrum, and the Federal Communications Commission approved a plan for a constellation of 15,000 satellites. RBC Capital Markets analysts said the moves improve Starlink Mobile’s ability to compete with the large carriers on coverage, and the firm carries a 225 dollar target, more than 50 percent above the roughly 148 dollars at which the shares traded on Tuesday morning. Cantor Fitzgerald’s Colin Canfield has a 246 dollar target and expects management’s commentary to center on connectivity and on how the company partners or competes its way into other markets.
Those targets describe a company valued on a forecast rather than on a record. At 148 dollars, the shares imply a market value that requires the AI business to keep compounding at triple-digit rates and the satellite business to convert subscribers into margin. The second quarter’s net loss, in a period of rapid revenue growth, is the arithmetic of a company that spends before it bills. Capital expenditure has risen in step with the compute buildout, and analysts will be listening for a figure rather than a direction.
Three questions will decide how the report is read. The first is whether management puts a number on 2027 capex, because the free cash flow profile of the whole company depends on it. The second is how much of the AI revenue is contracted for multiple years versus billed monthly, which determines whether the segment behaves like a cloud business or a chip reseller. The third is Cursor: whether the acquisition adds software revenue on top of compute demand, or simply moves existing customers onto a platform SpaceX now owns.
There is a credit dimension that equity analysts have started to raise. S&P has flagged the financing risk embedded in the AI buildout, and SpaceX funds compute the way its peers do, with debt against contracts. Yields on the bonds of smaller AI cloud companies have moved into distressed territory this month, and a company of SpaceX’s size is not immune to the question of what its marginal cost of capital has become.
The shares have traded above the IPO price since June and the sell-side is broadly positive, which sets up a specific kind of risk. A quarter that beats on revenue and misses on margin will be tolerated. One that shows AI growth decelerating from 146 percent, or capex guidance above what the models assume, will not. The stock has no history of drawdowns to show how it behaves when the narrative slips.
What the company says about Starlink’s subscriber additions deserves the closest reading of the operating numbers. A satellite network that is adding subscribers slowly while spending on spectrum and a 15,000-satellite constellation is a business in a transition, and transitions are expensive before they are profitable. The AI business may be growing faster, but it is also the one whose customers can move their workloads to whichever provider has capacity. A satellite in orbit cannot be replicated by a competitor next quarter, which is why the slower-growing segment is the more durable one.
SpaceX has spent a decade being valued on what it might do. It is now being valued on what it reports every three months, and the reports have to carry both stories at once. The results due after the close are the first real test of whether the market will pay for the transition or only for the growth.
