Goldman Sachs has presented prospective investors with projections showing SpaceX’s artificial intelligence revenues climbing from $3.2 billion in 2025 to $322 billion by 2030, according to a Financial Times report citing materials shared directly with a prospective backer. The figures form the central pillar of a $1.78 trillion target valuation for SpaceX’s planned initial public offering, which Goldman Sachs is leading – a structural conflict that analysts and investors have been quick to flag. The bank is simultaneously the architect of the growth thesis and the party with the most direct financial interest in seeing investors accept it.
The AI projections build in an aggressive near-term ramp. Goldman forecasts SpaceX’s AI revenues reaching $15.6 billion in 2026, a 388% jump from the 2025 base, then climbing to $34.5 billion by 2027 before accelerating sharply through the end of the decade. Total company revenues are projected to reach $474 billion by 2030, up from $18.7 billion last year. Underlying earnings are forecast to expand from $6.6 billion to $352 billion over the same period, with free cash flow swinging from negative $13.8 billion last year to positive $72 billion by 2031.
The projections require SpaceX’s AI unit – built around the Grok model ecosystem, previously operated under the xAI banner – to not only close the gap with established AI players but to surpass Anthropic, Google, and OpenAI across software development, autonomous agents, and enterprise interfaces. The entire valuation rests on an assumed total addressable market of $26.5 trillion for AI, dwarfing the $2 trillion market attributed to Starlink and rocket launches combined. Even within SpaceX’s own projected 2030 revenue mix, AI at $322 billion would outpace Starlink’s forecast $144 billion by more than two to one, and the core launch business at $8.3 billion by nearly forty to one.
Goldman beat out Morgan Stanley, JPMorgan, Citigroup, Bank of America, and UBS to secure the lead underwriter role – an arrangement that will generate tens of millions of dollars in fees across the participating banks, with Goldman capturing the largest share.
The Forecast Under Pressure
The commercial foundations of the AI projection face direct challenges from the unit’s own recent history. The Grok-based AI business lost $6.4 billion last year and holds a small fraction of the consumer and enterprise user base that the revenue model would require. Musk’s track record in managing the operation has added to investor uncertainty: all ten founding partners of the AI unit departed within its first twenty-four months. The gap between current performance and the 2030 target is not incremental – it requires the unit to establish dominant market positions in segments currently controlled by better-capitalized and more established competitors.
A detail buried in the source reporting underscores the present state of the business: SpaceX is currently leasing its 300-megawatt Colossus 1 data center in Memphis, Tennessee, to Anthropic – one of the competitors it must overtake to validate Goldman’s projections – in order to avoid leaving the facility idle.
The IPO in Context
The SpaceX offering is shaping up as one of the largest and most contested public listings in U.S. history. The $1.78 trillion target valuation would place it among the most valuable companies in the world on day one. Selling that valuation to institutional investors depends almost entirely on whether the AI revenue trajectory is credible – which is precisely the number Goldman, as lead underwriter, has an incentive to make compelling. The IPO roadshow is underway, with global asset managers being approached for commitments ahead of the listing. The deal structure reportedly aims to raise up to $86 billion in proceeds.
The convergence of factors – a wildly ambitious AI forecast, a lead bank with clear fee incentives, a business unit with a loss-making track record, and a valuation built on a $26.5 trillion market assumption – gives investors unusually clear lines of tension to work through before committing capital. Whether the roadshow resolves that tension or deepens it will likely determine the listing’s reception more than any individual data point in the Goldman deck.
