An enormous portion of revenue for cloud providers and other AI infrastructure players comes from just OpenAI and Anthropic. This extreme customer concentration creates a fragile system where the failure of one or two key players could trigger a widespread collapse.

While Starlink's satellite internet business is growing, President Gwynne Shotwell reveals that AI compute rental has become a massive and embarrassing profitable business for SpaceX. The insatiable demand for training hardware has turned 'Elon Web Services' into a key cash generator, overshadowing other revenue lines.

SpaceX's inaugural earnings report reveals a major business model evolution. The company's connectivity (Starlink) and AI segments now generate more revenue than its foundational space launch business. This highlights a successful pivot to higher-margin, scalable services over hardware-intensive launches.

Highlighting an extremely aggressive reinvestment strategy, SpaceX's AI segment spent $12.7 billion on capital expenditures in 2025 against revenues of only $3.2 billion. This massive bet on future growth is funded by its record-breaking IPO and other cash-generating business segments.

Nick and Jack reported that 84% of SpaceX's $7.6 billion in AI revenue came from renting out data centers rather than developing software. Anthropic alone pays $15 billion annually for compute rent, showing infrastructure providers are monetizing the AI boom far faster than model creators.

Elon Musk's xAI initially appeared to lag, but the real victory was building massive supercomputers. By leasing this compute power to rivals like Google and Anthropic, SpaceX's 'Neo Cloud' became its primary revenue source, revealing a highly profitable infrastructure-first AI strategy.

The massive three-year, $45 billion deal for Anthropic to use SpaceX's Colossus data centers instantly transforms SpaceX's revenue streams. This single contract makes the AI compute division a larger revenue generator than Starlink, signaling a strategic pivot for Elon Musk's company into a primary 'Compute as a Service' provider for the AI industry.

SpaceX's quarterly AI-related capital expenditures of $15.8 billion accounted for 86% of its total CapEx. This reveals that building competitive AI compute infrastructure now requires more capital than funding a fully-fledged private space exploration program, highlighting the staggering costs of the AI race.

SpaceX's spending on chips and data centers to power xAI is 50% more than the capital expenditure for its rocket and satellite divisions combined. This highlights a significant shift in deep tech, where the cost of computational infrastructure can now surpass that of complex, heavy industrial hardware.