SpaceX reports steep AI-related capital spending as shares fall after earnings
SpaceX disclosed a $18.4 billion capital outlay, over 80% of which was for artificial-intelligence infrastructure, and said the investment could be recouped within a year, but the stock slipped 7.5% in after-hours trading.
SpaceX said its second-quarter revenue rose 92% year-over-year, yet capital expenditures jumped to $18.4 billion, more than double the quarter’s sales. The company indicated that over 80% of the spending was directed to AI compute capacity, a sector where it trails established providers such as OpenAI, Anthropic and Google.
The outlay exceeded analysts’ average forecast of $13.22 billion, according to FactSet, and the after-hours decline erased earlier gains, leaving the shares more than 20% below the price at the June 12 IPO debut.
"We have been very efficient to date and I think we'll continue to be," CFO Bret Johnsen said on the earnings call. "On the AI compute side, we're able to deploy capital in such a way that we're getting less than a one-year payback."
SpaceX cited recent contracts that could generate up to $920 million a month from Google, up to $1.25 billion a month from Anthropic for three years, and up to $150 million a month from Reflection AI. Johnsen added that the company expects $6.7 billion of cloud-services revenue over a six-month period beginning in October, projecting annualized recurring revenue of $100 billion by year-end, assuming the completion of a $60 billion acquisition.
CEO Elon Musk reinforced the revenue target, stating that the $100 billion figure is “not a question mark” and would be achieved even without additional actions.
The AI unit, now operating as SpaceXAI, posted $2.56 billion in revenue for the quarter but recorded a $1.26 billion operating loss, following a $2.47 billion loss in the prior quarter. The company also disclosed a $354 million accrual for probable litigation costs related to emissions permits at its Memphis data center.
Analysts noted that SpaceX’s strategy of reselling AI compute capacity differs from its broader aerospace ambitions, and the firm’s ability to convert the heavy capital spend into sustainable profit remains under scrutiny.