Is SPCX Stock a Good Investment? SpaceX Earnings, Valuation and Key Risks

SpaceX reported third-quarter 2026 revenue of $14.58bn after the close on 20 October, up roughly 161% year on year and well clear of a consensus that sat between $12.7bn and $12.9bn, with net income of about $4.4bn against a loss of roughly $1.05bn in the same quarter of 2025. Adjusted EBITDA was $8.8bn, up 306%. On the numbers alone it was the quarter that turned a capital-consuming project into a profitable one at scale, and the shares, which closed the regular session at $147.79, rose in after-hours trading. For anyone asking whether SPCX stock is a good investment, that is the case for the company stated as plainly as it can be: triple-digit revenue growth, a genuine profit, and two businesses — satellite connectivity and AI compute — each large enough to matter on its own.

What would have to go right is equally plain, and it is a spending question rather than a demand question. Third-quarter capital expenditure was $19.6bn — more than the quarter's entire revenue — of which $16.4bn went into the Super Intelligence segment. On the earnings call, chief financial officer Bret Johnsen guided fourth-quarter capital spending to $35bn to $40bn — roughly double the $19.6bn the company had just spent, and a reversal of the guidance given in August, when management said capital spending over the following two quarters would run at a level similar to the second quarter's $18.4bn. At a market capitalisation in the region of $2tn, the valuation is not asking whether SpaceX can grow. It is asking whether a company spending more than it earns, at an accelerating rate, converts that into durable returns before the funding environment changes its mind. Everything below is an attempt to separate those two questions. See the full factor breakdown on the SPCX equity page.

What SpaceX actually reported

The quarter had two engines, and the newer one is now the bigger.

Connectivity — the Starlink business — generated about $5.2bn of revenue, up 69% from roughly $3.1bn a year earlier and ahead of analyst expectations near $4.84bn. The split was approximately $3.12bn from consumer Starlink subscribers and $2.13bn from government and enterprise customers. Starlink added 1.6 million consumer subscribers in the quarter to reach 13.6 million broadband consumer subscribers, a figure management noted excludes several hundred thousand enterprise subscribers and aviation customers. Monthly average revenue per user rose 5% sequentially to $69, helped by customer mix.

That combination — subscriber additions of 1.6 million in a single quarter and ARPU rising rather than falling — is unusual. Consumer broadband businesses normally buy volume with price. Starlink added volume at a higher price, which is what a capacity-constrained network looks like when the constraint eases.

Super Intelligence, the AI compute and software segment, was the larger contributor at about $7.6bn of revenue, up roughly 195% sequentially and some $6.7bn higher than a year earlier. The growth came principally from new cloud services agreements and from Grok usage, with the Cursor acquisition contributing as well. A segment that was immaterial twelve months ago is now more than half of group revenue.

Against that, the capital intensity. Capital expenditure of $19.6bn in the quarter, $16.4bn of it in Super Intelligence, exceeded revenue by roughly $5bn. The market's immediate reading was mildly favourable, because the quarter's spending landed in line with the guide given in August. The guidance is what reset the frame: $35bn to $40bn in the fourth quarter, attributed by management to added capacity coming online. Johnsen also guided first-quarter 2027 revenue to $33bn to $37bn, against an analyst average near $22.2bn.

Why the market reacted the way it did

The after-hours reaction was positive but the size of it is genuinely contested across data providers, with reported moves ranging from under 2% to around 8%. We are not going to pretend to a precise figure. What is clear is direction, and the direction is explained by the Q1 2027 revenue guide rather than by the Q3 print.

A $33bn to $37bn revenue guide for a single quarter, set against a $22.2bn consensus, implies the company believes the compute capacity it is building is already contracted or close to it. That is the only reading under which a doubling of quarterly capital spending is received as good news rather than bad: spending accelerates because demand is visible, not speculative. Investors who accept the guide treat the capex line as pre-sold capacity. Investors who do not treat it as the largest unhedged bet in the listed market.

The connectivity result supports the optimistic reading at the margin. A 69% revenue increase with ARPU up and 1.6 million net additions is evidence that SpaceX's existing capital programme — the launch cadence and the satellite constellation — has produced a business that monetises reliably. The argument for the compute build is, in essence, that the company has done this once already.

The Openbook read

Growth is as strong as this factor gets, and it is strong on two independent axes. Group revenue up 161%, connectivity up 69% against a mature-ish subscriber base, and a segment that went from nothing to $7.6bn in a year. The Q1 2027 guide, if met, extends it rather than annualising it away. There is no sensible reading of this release in which Growth is anything other than the dominant factor.

Profitability has improved more than any single quarter usually allows. A swing from a $1.05bn loss to $4.4bn of net income, with adjusted EBITDA at $8.8bn and up 306%, is a step change rather than a trend. Two cautions belong in the score. First, adjusted EBITDA is the metric the company leads with, and in a business capitalising this much spending, EBITDA flatters the economics by construction — depreciation on $19.6bn of quarterly capex is a real cost arriving later. Second, sources disagree on reported earnings per share for the quarter, with estimates clustering between $0.10 and $0.15 against a reported figure near $0.35, so we are not treating EPS as load-bearing here.

Momentum is positive but is the factor most exposed to the capex narrative. The stock has been trading on the AI compute story, and the fourth-quarter spending guide has made that story more binary, not less. Momentum scores built on a narrative that can be falsified by a single contract disclosure carry more fragility than the price chart suggests.

Solvency is the factor that deserves the most attention and gets the least coverage. Capital expenditure that exceeds revenue is sustainable only while funding is available on acceptable terms, and a guided $35bn to $40bn of fourth-quarter spending against roughly $14.6bn of quarterly revenue means the gap is widening rather than closing. The company's own profitability helps, but $8.8bn of adjusted EBITDA does not fund $35bn of quarterly capex. Something else does: cash on hand, new debt, or equity. Until the funding mix for the fourth quarter and 2027 is disclosed, Solvency is the factor carrying the uncertainty, and it is a structurally different kind of risk from the one a connectivity-only SpaceX would have had.

Reward/Risk is where the roughly $2tn market capitalisation does its work. At that size the company is already priced as one of the largest listed businesses in the world, which means the reward case requires the Q1 2027 guide to be not just met but extended, repeatedly. The risk case does not require anything to go wrong operationally. It only requires the compute demand curve to flatten while the spending commitments remain fixed — the classic infrastructure mismatch, where capacity arrives on a two-year lead time and demand revises on a two-quarter one. The asymmetry runs in both directions and is wider than the valuation of a profitable, fast-growing company would normally imply.

The read-across

The clearest read-across is to the AI compute supply chain. A fourth-quarter capital budget of $35bn to $40bn from a single buyer is a material addition to sector demand, and it lands on the same suppliers serving the established hyperscalers. Nvidia is the obvious beneficiary, and the power, cooling and data-centre construction names sit behind it. Investors tracking the aggregate build can compare capital intensity across the group on the screener; what SpaceX has done is add a buyer whose spending trajectory is steeper than anyone else's from a standing start.

For the hyperscalers themselves, the read is more ambiguous. Microsoft and its peers have spent two years persuading investors that enormous AI capital budgets are disciplined and demand-led. A new entrant guiding to a near-doubling of quarterly spend validates the demand thesis and simultaneously raises the question of whether the market is building more capacity than it can fill. Both effects are real and they point in opposite directions for sentiment.

In satellite connectivity, the Starlink numbers are the competitive datapoint. ARPU rising to $69 while adding 1.6 million subscribers tells every other operator that the pricing umbrella in consumer satellite broadband is not coming down, and that the subscriber ceiling is higher than the sceptics modelled. For the launch market, a profitable SpaceX funding its own compute build is a competitor with a lower cost of capital than it had a year ago.

What to watch next

• Fourth-quarter results, expected late January or February 2027: whether capital expenditure lands inside the guided $35bn to $40bn, and what the funding mix was. This is the single most informative disclosure ahead.
• The Q1 2027 revenue guide of $33bn to $37bn: it implies roughly a doubling from the third quarter in two quarters. Any softening of that range, or a shift in how much of it is contracted, changes the capex case directly.
• Super Intelligence segment disclosure: the segment is now more than half of revenue and was built partly by acquisition. Clearer separation of organic growth, contracted backlog and acquired revenue would materially improve the quality of the Growth score.
• Starlink subscriber additions and ARPU: 1.6 million and $69 are the benchmarks. ARPU is the more sensitive of the two, because it is the evidence that the network is not yet having to buy growth.
• Depreciation: as the 2026 capital programme enters service, the gap between adjusted EBITDA and net income will widen. The first quarter in which that is visible is the real test of the Profitability score.

This was, unambiguously, a strong quarter. The question it leaves is not whether SpaceX can grow but whether a company that spent $19.6bn to earn $14.58bn, and plans to spend roughly twice as much again in the next three months, is being valued on the capacity it is building or on the returns that capacity has yet to produce.