Elon Musk's SpaceX IPO Friday values the company at $1.75 trillion, making it the largest public offering in history. But the headline number masks what investors actually back: not rocket launches, but an artificial intelligence infrastructure play that positions SpaceX as a data-center operator in orbit.
The company's AI division lost $6.4 billion last year. That loss reflects Musk's stated goal of deploying a million satellites to create a global data-center network in low Earth orbit. SpaceX plans to offer compute capacity and low-latency connections that could serve AI training and inference workloads. The satellite constellation, initially designed for broadband, transforms into a distributed computing platform.
The valuation tells the real story. SpaceX's worth has more than doubled since December, driven less by launch cadence improvements and more by this orbital infrastructure thesis. Investors price in the possibility of SpaceX capturing a slice of the trillion-dollar AI compute market by offering advantages traditional data centers cannot: global coverage, latency measured in milliseconds, and infrastructure Musk controls entirely.
Compare this to Apple's opposite gamble. Apple rejected building its own AI infrastructure at massive scale. Instead, the company partnered with OpenAI and other providers, integrating their models into software. Apple avoided the capital expenditure and operational complexity that SpaceX embraces. Apple bets on software and integration. Musk bets on physics and infrastructure.
The timing matters. Cloud providers compete fiercely on compute costs and speed. Starlink's existing satellite network, deployed and operating, gives SpaceX a head start building its orbital data-center layer. The company can monetize existing infrastructure while expanding capacity.
Separately, AI television launches this week. The trend reflects broader adoption of generative models in consumer products beyond chat interfaces. TV platforms experiment with AI-generated