SpaceX’s First Public Earnings Statement Shows The Financials Of An AI Company In 2026


That’s a lot of capital expenditure.

SpaceX has published its first earnings report since going public on June 12. The company’s Q2 topline is broadly positive. As a whole, SpaceX increased revenue by 92 percent year-over-year, bringing its total revenue for the quarter to $7.8 billion. The company also shaved the gap between its revenue and operating expenses by more than half, reducing net losses to $541 million, down from $1 billion a year ago. But digging deeper, the financials reveal just how expensive is to run an AI company. 

Over the past three months, the division previously known as xAI generated $2.56 billion in revenue, an increase of 247 percent from $737 million one year ago. SpaceX attributes that increase primarily to “Cloud Services Agreements” like the highly publicized deals with Anthropic and Google to use its data centers. The company also managed to cut its operating loss to $1.25 billion, down from $1.54 billion. However, SpaceX significantly ramped up CapEx spending. Year over year, that number increased by a whopping 2,013 percent to $15.8 billion for the quarter. For context, the rocket launch part of SpaceX’s business by comparison spent a paltry $1.17 on maintaining and improving its fixed assets.

While we had an idea of just how expensive it was to run an AI company thanks to reporting from people like Ed Zitron, SpaceX’s financials give us our first definitive look at the inside of an AI startup. And keep in mind this is all coming from the company that many consider the fourth horse in the AI race behind Anthropic, Google and OpenAI. Further complicating those numbers is that Anthropic has agreed to pay $1.25 billion per month through May 2029 to use SpaceX’s Colossus 1 data center for its compute needs. It will be interesting to see how SpaceX’s Cursor acquisition affects the AI division’s bottom line when that deal closes sometime this quarter. 



Source link

  • Related Posts

    Open-weight AI models are catching up to the frontier. The safety gap remains. 

    As policymakers debate how to govern increasingly powerful AI systems like OpenAI’s GPT-5.6 Sol and Anthropic’s Mythos, a Chinese open-weight model has narrowed the gap with the industry’s leaders.   GLM-5.2,…

    OpenAI And Anthropic Models Went On A Hacking Spree When Tested By The UK’s AI Research Institute

    Both OpenAI and Anthropic recently admitted that their models escaped from their testing environments and hacked into outside organizations on their own. Now, the UK’s AI Security Institute (AISI) has released…

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You Missed

    N.S., company suing each other over care for children, disabled persons

    N.S., company suing each other over care for children, disabled persons

    There are a few gooduns in August’s Xbox Game Pass line-up, including a surprisingly sexy, and funny, furniture dating sim

    There are a few gooduns in August’s Xbox Game Pass line-up, including a surprisingly sexy, and funny, furniture dating sim

    Two Cities With a Tangled History

    Two Cities With a Tangled History

    PM Carney teams up with TO mayor Olivia Chow to highlight fresh support for affordable housing and tour ‘public transit site’ with Ford

    PM Carney teams up with TO mayor Olivia Chow to highlight fresh support for affordable housing and tour ‘public transit site’ with Ford

    Surrey taxpayers could face huge bill to accommodate SSMUH

    Surrey taxpayers could face huge bill to accommodate SSMUH

    This federal judge ruled against Trump. Then the threats began.

    This federal judge ruled against Trump. Then the threats began.