The pitch behind the largest U.S. stock market debut in history rests on one number, and it asks investors to suspend a lot of disbelief. Goldman Sachs (NYSE:GS | GS Price Prediction), the lead underwriter on the deal, told a potential IPO investor it expects SpaceX’s AI division revenue to climb roughly 100 times by 2030.
That projection, reported by the Financial Times, is the central scaffolding holding up the $1.78 trillion valuation SpaceX is pitching during its current roadshow. The deal could raise as much as $86 billion.
For investors watching from the sidelines, the story is undeniably exciting. The deeper question is whether the math behind the story can survive contact with reality.
The Number
Goldman Sachs expects SpaceX’s AI division revenue to rise to $322 billion by 2030 from $3.2 billion in 2025. Reportedly, that’s the figure underwriting the trillion-dollar pitch.
Goldman beat out Morgan Stanley, JPMorgan, Citigroup, Bank of America, and UBS to land the lead role on the IPO. Goldman declined to comment on the matter, and SpaceX didn’t respond to a request for comment, according to the Financial Times.
What It Means
The AI ramp is only the beginning of the story. Goldman expects the AI segment to rise 388% to $15.6 billion in 2026 compared with the prior year, and to reach $34.5 billion in 2027. SpaceX’s total revenue is anticipated to reach $474 billion in 2030, up from $18.7 billion last year.
Starlink revenue is projected to hit $144 billion in 2030, less than half of what the AI unit is forecast to generate, and the rocket division is expected to produce $8.3 billion in 2030, up from $4.1 billion last year. Moreover, SpaceX’s adjusted EBITDA is forecast to climb to $352 billion in 2030 from $6.6 billion in 2025. Free cash flow, which ran negative $13.8 billion last year, is expected to swing to a positive $72 billion in 2031.
The operational backdrop, per the S-1, includes roughly 1.3 billion accounts active in the trailing 12 months ended March and roughly 550 million monthly active AI users across Grok and X. The SpaceX filing also describes an ambitious vision to launch 100 gigawatts of compute to space each year via AI compute satellites, a space-based compute-as-a-service concept that the bull case effectively requires.
Market Reaction
SpaceX is privately held, so there’s no share price to track. The closest real-time read on investor conviction comes from the prediction markets, where Polymarket is currently pricing a 98% probability of an IPO by June 30.
That probability climbed from 95% on May 29 to 98% on June 4, suggesting traders see the listing as a near certainty in the next several weeks. SpaceX plans a NASDAQ listing under ticker SPCX as early as June 12, with pricing set at $135 per share on June 3.
Bear Case
The bears have plenty of ammunition, and most of it sits inside the AI unit itself. xAI incurred a loss of $6.4 billion in 2025, and for the AI growth forecast to be justified, SpaceX’s Grok family of models would have to catch up to and surpass more advanced labs including Anthropic, Alphabet‘s Google (NASDAQ:GOOGL), and OpenAI in coding, cybersecurity, agents, and chatbots.
The operational track record doesn’t help. Tesla (NASDAQ:TSLA) CEO Elon Musk pushed out all 10 of his xAI co-founders within two years, and the unit has delivered subpar performance, leaving it with only a fraction of the consumer and corporate subscriptions needed to generate revenues. Musk rented out the Colossus 1 data center to rival Anthropic, as the 300-megawatt facility in Memphis, Tennessee fell underutilized after Grok failed to gain traction.
Then there’s the assumption underpinning the entire pitch. The valuation rests on a claim, per SpaceX’s IPO prospectus, that the xAI division has a total addressable market of $26.5 trillion, which dwarfs the roughly $2 trillion total addressable market outlined for Starlink and space operations. However, an addressable market is not a revenue line, and the gap between the two is where this thesis can break.
Competitively, Google, along with Anthropic and OpenAI, currently sets the bar Grok needs to clear. The Financial Times frames Goldman’s projections as aggressive bets underpinning the broader AI investment boom that has pushed U.S. stocks to record highs.
Bottom Line
The 100x AI revenue projection functions as aggressive guidance. It serves as the load-bearing wall of a trillion-dollar valuation, and it asks a unit that lost $6.4 billion last year to dethrone the most advanced AI labs in the world while simultaneously commercializing space-based compute.
Catalysts to watch include the IPO roadshow and final pricing, and whether Grok closes the gap with leading AI labs. It’s also a good idea to check SpaceX’s AI segment revenue milestones against the $15.6 billion target for 2026, and the company’s path from negative $13.8 billion free cash flow toward the $72 billion positive figure Goldman pencils in for 2031. Each milestone either validates the model or quietly erodes it.
The bears may concede that SpaceX can grow. Their argument is that the price tag already assumes it grows faster than any company in history, in a segment where it currently trails. That’s the bet on the table, and it deserves to be sized accordingly.
