
s the core AI bet - but it's described as early-stage and heavily loss-making, with a 2025 operating loss of $6.4 billion. That spending is accelerating: SpaceX reported $20.7 billion of capital expenditures in 2025, including $12.7 billion aimed at AI.
Why should I care?
For markets: A make-or-break test for the AI premium.
A $75 billion raise would be enormous for any listing, and it would force investors to choose between two ways of valuing SpaceX: current earnings power from Starlink versus a big, expensive AI buildout. The key question is whether markets keep rewarding "AI-first" narratives when the bill is already visible - xAI's projected 2025 operating loss is larger than Starlink's operating profit. It could also set a benchmark for other AI companies that may eventually go public, making comparisons on growth, spending, and profitability harder to dodge.
Zooming out: Enterprise AI is turning into a distribution game.
The filing suggests the next stage of enterprise AI won't just be about having the smartest model - it'll be about getting software embedded inside big organizations. SpaceX says it plans a specialized enterprise sales team and "forward deployed engineers", meaning staff who work directly inside customers to help roll out tools and retrain workflows. It also hints at a wider stack like Grok Enterprise and more infrastructure investment, showing how the race is spreading from software into the computing capacity needed to run it.