Goldman Sachs assigned SpaceX's artificial intelligence operations a $322 billion revenue target for 2030, positioning compute services as 68% of total company sales in a projection that reframes the firm's Nasdaq IPO narrative from launch provider to infrastructure lessor. The estimate arrives as SpaceX maintains its $350 billion private valuation and begins disclosing division-level financials to institutional allocators ahead of an anticipated 2026 public offering.
The projection assumes SpaceX converts Starlink's 7,000 operational satellites into distributed compute nodes serving hyperscalers and AI training workloads, monetizing latency advantages and orbital positioning that terrestrial data centers cannot replicate. Goldman's model prices SpaceX AI capacity at $4.20 per GPU-hour by 2028, a 31% premium to AWS equivalents, with margins approaching 54% as capital costs amortize across dual-use satellite hardware already deployed for broadband service. The math requires SpaceX to capture 18% of the global AI infrastructure market within six years, comparable to Microsoft Azure's current enterprise share.
The thesis matters because it separates SpaceX from aerospace comps and embeds the company in data center and cloud multiple territory where forward revenue growth commands 12x-18x EBITDA versus the 6x-9x range typical for launch contractors. If Goldman's numbers hold, SpaceX's AI segment alone would exceed the combined 2030 revenue of Lockheed Martin and Northrop Grumman, creating a valuation floor divorced from launch cadence or Starship development timelines. The projection also implies SpaceX has solved latency and thermal dissipation constraints that currently limit orbital compute density to research workloads, a technical claim the company has not yet demonstrated at commercial scale.
Allocators should watch for two disclosure events: first, SpaceX's submission of AI revenue breakouts in pre-IPO roadshow materials expected in Q1 2026, which will confirm whether the division already generates material cash flow or remains a projection; second, Amazon's response, given Project Kuiper's 3,400-satellite constellation could replicate the same infrastructure model and compress SpaceX's pricing power before 2028. Goldman did not specify how much of the $322B forecast depends on exclusive contracts versus open-market spot pricing.
The 68% sales mix implies SpaceX's launch business plateaus near $150 billion annually by decade-end, roughly 4x current run-rate despite Starship entering full service.