Elon Musk's SpaceX IPO prices the company at $1.75 trillion, making it the largest initial public offering in history. The headline number masks what investors are actually buying: a sprawling bet on AI infrastructure and orbital data centers, not rockets alone.
SpaceX's AI division burned through $6.4 billion in losses last year. The unit focuses on developing custom silicon and machine learning systems that Musk plans to distribute across a constellation of data-center satellites. The company intends to deploy roughly one million satellites into orbit, creating a distributed computing network accessible from anywhere on Earth. This represents a fundamental shift from SpaceX's traditional launch services business.
The valuation has more than doubled since December, jumping from roughly $750 billion to $1.75 trillion. That acceleration reflects investor appetite for infrastructure plays tied to AI workload distribution, not improvements in Starship or traditional launch economics.
The satellite-based AI compute bet operates on a simple premise: ground-based data centers face congestion and latency constraints. Orbital infrastructure sidesteps both problems by placing processing power directly in space. However, this strategy carries massive technical and regulatory hurdles. Deploying one million satellites requires unprecedented coordination with telecommunications regulators, space debris management, and laser-based inter-satellite linking at scale never attempted.
Apple's recent moves offer a contrasting datapoint. Rather than building proprietary infrastructure, Apple licensed cloud services from OpenAI and Google, keeping AI compute off its devices. Musk's approach goes opposite: vertically integrate satellites, ground stations, chips, and software into one controlled system.
The SpaceX IPO values the company's future more on orbital infrastructure than launch frequency. Wall Street is pricing in success for an AI satellite constellation that remains largely theoretical. If execution stumbles, the valuation rests on speculative future revenue from services that don