JPMorgan Raises SpaceX Price Target to $240
Claire Weston
JPMorgan raised its SpaceX price target from $225 to $240, driven by AI compute commercialization far ahead of schedule — the 2027 revenue forecast jumped 64% to $123 billion in a single revision, and the trillion-dollar revenue milestone moved up a full year.
Why such a massive revenue forecast revision in one go?
AI cloud contracts are landing fast: Anthropic at $920 million/month, Google at $1.25 billion/month (from October), Reflection AI at $150 million/month, plus an anonymous client paying $6.7 billion over six months.
This means → SpaceX's AI compute business is no longer a pitch deck — it is real cash at the multi-billion-dollar-per-month level.
JPMorgan accordingly raised 2026 revenue by 14% to $47 billion and 2027 revenue by 64% to $123 billion; annualized recurring revenue is expected to top $100 billion by December 2026.
Who is the mystery client?
$6.7 billion over six months, or over $1 billion a month — the list of buyers who can write that check is extremely short.
Market speculation centers on Meta, OpenAI, Microsoft, SoftBank, Nvidia, or similar heavyweights.
In plain terms = regardless of who it is, the contract itself proves one thing: compute scarcity is structural, and major buyers will lock in premium long-term deals to secure capacity.
Why could compute pricing be so much higher than modeled?
Musk believes Vera Rubin architecture — Nvidia's next-generation GPU platform — could command $30–50 per watt, far above JPMorgan's prior model range of $11–16 per watt.
This means → at the same compute scale, monetization could double or triple.
Higher pricing on top of greater scale could pull SpaceX's $100 billion-plus AI revenue target forward from 2028 to 2027.
Where does the money come from? How heavy is the capex burden?
SpaceX plans to expand compute deployment from over 2 GW in 2026 to 5–10 GW in 2027, well above JPMorgan's prior estimate of 4.2 GW.
The direct consequence: 2027 capex forecast jumped from $128 billion to $196 billion; the 2027 free-cash-flow gap is roughly $108.2 billion.
In plain terms = SpaceX is running a "burn-and-earn" model — new compute investments pay back in roughly one year or less, leading the industry, but cumulative negative free cash flow from 2026 to 2030 is projected above $300 billion, requiring continuous outside financing.
From ground to orbit — what is "Starmind"?
SpaceX has committed to exclusive use of Nvidia GPUs in its data centers and is co-developing the first AI satellite — named Starmind — with Nvidia, also built on the Vera Rubin architecture.
Launches are planned to begin in 2027.
This reflects an ambition that goes beyond terrestrial data centers — extending compute into orbit effectively merges SpaceX's rocket business with its AI business.
What is the biggest near-term risk?
911.5 million shares come off lockup on August 6 — potential supply equal to 143% of the current 639 million share float. SpaceX fell 13.6% on Wednesday to $108.27.
JPMorgan believes the market has already adjusted positioning to some degree, but multiple additional unlock tranches follow in the months ahead.
Longer-term risks are also significant: Starship delays, GPU and power supply constraints, over $300 billion in cumulative negative free cash flow dependent on financing, export-control and spectrum-allocation uncertainty, and governance concentration from Musk's roughly 82% voting control.
Content is for reference only, not financial advice.