SpaceX has discovered a lucrative new business model: renting artificial intelligence computing power to the companies building the future of AI. Through its acquisition of Elon Musk's xAI startup earlier this year, SpaceX now operates massive data centers like Colossus and Colossus II, which it leases to major AI developers. The rental agreements alone generate more monthly revenue than SpaceX's entire 2025 earnings, fundamentally reshaping the company's financial picture just weeks after its June public offering (Source 1, 2, 3).

Why Are AI Companies Paying Billions to Rent Computing Power?

The AI industry faces an unprecedented bottleneck. Major AI companies need enormous amounts of computing capacity to train and run their large language models, but the global supply of specialized chips, power infrastructure, and data center space cannot keep up with demand. Anthropic, which develops the Claude AI model, agreed to pay SpaceX $1.25 billion per month for three years to access more than 220,000 Nvidia graphics processing units (GPUs) and over 300 megawatts of computing power at the Colossus 1 facility in Memphis, Tennessee. Google signed a separate deal weeks later, committing to pay $920 million monthly starting in October for roughly 110,000 GPUs through June 2029.

Combined, these two contracts alone are worth $2.17 billion per month, or approximately $26 billion annually. To put this in perspective, SpaceX generated just $18.67 billion in total revenue during 2025. A single tenant's monthly rent now exceeds what the entire company earned in a year.

How SpaceX Built Its AI Infrastructure Advantage

When SpaceX acquired xAI in February through an all-stock deal valued at roughly $1.25 trillion, it inherited not just Musk's AI ambitions but also the physical infrastructure already built to support them. The company now operates data centers equipped with hundreds of thousands of specialized AI chips from suppliers like Nvidia and Advanced Micro Devices. These facilities were originally constructed to train Grok, xAI's AI model, but SpaceX quickly realized it could monetize excess capacity by renting it to competitors (Source 1, 2).

The rental model works because both Anthropic and Google face urgent capacity constraints. Anthropic needs the infrastructure to support its Claude models, while Google requires what it calls "bridge capacity" for Gemini Enterprise, its agentic AI platform, which has grown faster than Google's own internal infrastructure can handle. Regardless of which AI model ultimately dominates the market, SpaceX profits as the landlord.

What Makes SpaceX's Long-Term AI Strategy Unique?

SpaceX's ambitions extend far beyond Earth-based data centers. The company has filed an application with the U.S. Federal Communications Commission to launch and operate up to one million satellites as part of an orbital data center system internally called Starmind. These AI satellites, designed with 70-meter wingspans and 150-kilowatt peak compute payloads, would run on uninterrupted solar power and require minimal cooling, addressing the two largest operating expenses for traditional data centers. Musk has stated that space represents "the only way to scale AI," and the company plans to begin launching these satellites by 2028 using its Starship rocket.

Nvidia has already developed specialized hardware for this vision. The Space-1 Vera Rubin Module, announced at GTC 2026, is designed to deliver data center-class AI computing in space, delivering up to 25 times the AI compute power per GPU compared with the H100 chip. The module is expected to be available in 2027.

How Does This AI Revenue Reshape SpaceX's Business?

• Connectivity Shift: Starlink, SpaceX's satellite internet business, generated $11.4 billion in 2025 and was the company's largest revenue source. The AI rental contracts will generate more than double that amount annually from just two customers.
• Revenue Acceleration: Wall Street forecasts SpaceX will more than double its total revenue to $38.8 billion in 2026 and reach $72.4 billion in 2027, driven largely by AI infrastructure deals (Source 1, 2).
• Contract Terms and Risk: Google's lease includes a 90-day exit clause after December, while Anthropic's ramp-up period has already passed, meaning both companies could theoretically reduce or terminate their commitments, though the financial incentives currently favor continuation.

What Are the Financial Implications for SpaceX Investors?

Despite the explosive growth in AI revenue, SpaceX's stock valuation remains controversial. The company went public on June 12 at $1.77 trillion and surged to a peak market capitalization of $2.9 trillion before declining 35 percent to close at $145 per share on July 10. Based on trailing 12-month revenue, SpaceX trades at a price-to-sales ratio of 98.9, making it 15 times more expensive than the Nasdaq-100 index, which has a price-to-sales ratio of 6.4 (Source 1, 2).

Even using Wall Street's optimistic 2027 revenue forecast of $72.4 billion, SpaceX's forward price-to-sales ratio remains 26.3, significantly higher than comparable technology companies. Financial analysts have suggested the stock could decline 50 percent or more if the company fails to meet growth expectations, though the AI infrastructure revenue provides a more stable, recurring income stream than SpaceX's traditional launch business (Source 1, 2).

What Does This Mean for the Broader AI Industry?

SpaceX's success in monetizing AI infrastructure highlights a critical reality in the artificial intelligence market: compute capacity has become a scarce resource commanding premium prices. The company has positioned itself as an essential intermediary between chip manufacturers and AI developers, capturing significant value regardless of which AI models ultimately win market share. Whether Claude, Gemini, or Grok becomes the dominant AI platform, SpaceX profits from providing the underlying infrastructure.

The long-term vision of orbital data centers, while ambitious and uncertain in execution timeline, represents SpaceX's attempt to solve the fundamental constraints limiting AI scaling. If successful, it would give SpaceX an unprecedented advantage in a market that could eventually be worth trillions of dollars. For now, the company's Earth-based rental agreements have already transformed it from a space transportation company into a critical piece of AI infrastructure, generating revenue that dwarfs its traditional business lines.