Morgan Stanley: SpaceX AI Business Severely Undervalued, $300 Target with $600 Bull Case
Alina Collins
Morgan Stanley set a $300 base-case target for SpaceX and $600 in a bull scenario, arguing the market prices its AI business at just ~$12/share — a segment on track to clear $10 billion in annual revenue, yet valued below a typical cloud startup.
Why does AI show up at only $12 a share?
Morgan Stanley strips SpaceX into two halves: space-and-connectivity is worth roughly $127/share. Subtract that from the current ~$140 stock price, and the implied value of the entire AI business — consumer and enterprise — is about $12/share.
This means → the market assigns a 2028 EV/sales multiple of just over 1× to SpaceX's AI segment, below even early-stage cloud peers.
In plain terms = investors buying SpaceX today are getting the AI business almost for free. Morgan Stanley expects this discount to narrow as Cursor and Grok milestones keep landing.
What does the $60 billion Cursor deal actually buy?
In June 2026, SpaceX announced a $60 billion all-stock acquisition of Anysphere, Cursor's parent company, expected to close in Q3. The two sides had already begun collaborating in April, with SpaceX and xAI opening the Colossus data center — a massive AI compute cluster — for Cursor's model training.
Morgan Stanley lays out three strategic pillars: ① Compute leverage — Colossus cuts Cursor's dependence on third-party AI labs and compresses per-token costs. ② Data flywheel — Cursor holds real developer-usage data from 50,000 enterprises, covering 64% of the Fortune 500. ③ Enterprise beachhead — Cursor becomes the entry product for a broader enterprise AI platform.
The combined map Morgan Stanley sketches: Cursor as the application layer, Grok as intelligence, X as real-time data, SpaceX AI as compute, Starlink as global connectivity, and Tesla as the path to physical AI.
How fast is Cursor's revenue growing?
Cursor's ARR — annualized recurring revenue, a standard SaaS metric that projects current subscription run-rate over a year — is scaling at one of the fastest rates in software history: $100 million in January 2025, roughly $4 billion by June 2026, an ~8× increase in a single year. Enterprise B2B clients account for about 75%.
Morgan Stanley's ARR trajectory: ~$8 billion by end of 2026 → ~$17 billion in 2027 → ~$33 billion by 2030. Cursor is projected to contribute about 19% of SpaceX AI's total revenue in 2027.
This means → Cursor is no longer just a coding tool — it is becoming the core growth engine of SpaceX's AI revenue.
How does gross margin flip from negative to positive?
Cursor has historically burned cash on every subscription sold. Gross margin in Q1 2026 was roughly -23%, because each user session required paying third-party AI labs per-token inference fees — essentially a toll on every query.
Morgan Stanley expects gross margin to turn positive in Q3 2026 and reach the low-60% range by 2030. The lever: shifting inference workload from external models to in-house ones. The share handled by Cursor's own Composer models and Grok rises from ~33% in 2026 to ~72% by 2030.
In plain terms = Cursor was paying a "toll" to outside labs on every subscription sold; as it brings most of the work in-house, the toll disappears. The cost gap is stark: Grok 4.5 runs at ~$2.59 per task, top external frontier models at $7–8+; Cursor's Composer series costs just $0.04–0.08 per task.
How much room is left in enterprise AI spending?
Morgan Stanley cites the Ramp AI Index: as of June 2026, median monthly AI spend per U.S. employee was about $11, up 167% year-over-year. The top 1% of companies exceeded $4,880, growing ~240–250% YoY. Yet nearly half of U.S. companies still pay nothing for any AI tool.
In coding AI specifically, the median engineer now holds an AI coding-tool license, and top-quartile firms have coverage near 90%. But only about 28% of shipped code at the median company is AI-assisted.
This reflects a gap between "tools distributed" and "tools actually used." Morgan Stanley sees the gap closing fast. Whether Cursor can convert license coverage into real active usage — and the seat-and-revenue growth that follows — is the key validation point for the entire re-rating thesis.
Content is for reference only, not financial advice.