Analysts: SpaceX's Terrestrial AI Infrastructure Value Not Yet Fully Priced In
nashnova research
Bernstein and Deutsche Bank both raised SpaceX price targets to $268 and $250 on the same day, arguing the market undervalues its terrestrial AI compute business — implied upside ranges from 70% to 82%.
Two banks turned bullish at once — what did they see?
Bernstein maintained its outperform rating and raised its target from $248 to $268, implying roughly 82% upside from last Friday's close.
Deutsche Bank kept its buy rating and lifted its target from $230 to $250, implying about 70% upside.
This means → both firms share one core thesis: SpaceX is more than a rocket company, and its terrestrial AI compute business has not been properly priced in.
Terrestrial AI compute — why does this matter?
Bernstein analyst Douglas Harned noted that the market pins SpaceX's multi-trillion-dollar valuation logic on orbital data centres, yet terrestrial AI buildout is a critical complement to the overall revenue path.
In plain terms = everyone is watching "servers in space" while overlooking SpaceX's massive ground-level compute expansion.
Deutsche Bank analyst Edison Yu was more direct: SpaceX's potential to convert compute power into stronger inference and intelligence capabilities "continues to be under-appreciated."
Why is its compute cheaper than peers'?
SpaceX plans to add roughly 10 gigawatts of incremental compute capacity by the end of next year, priced at about $35 billion per gigawatt.
The industry benchmark sits around $50 billion per gigawatt — SpaceX is nearly 30% cheaper.
This means → once that cost gap scales, it becomes a competitive moat: same compute, lower price to win customers, and fatter margins.
Where does broader Wall Street stand?
Of the 43 analysts covering SpaceX, 33 rate it buy or strong buy — the bull case is the consensus view.
Yet the stock has been flat over the past month, suggesting the market is waiting for harder evidence.
This reflects a tension: analysts are collectively bullish, but the share price hasn't moved — the "under-priced" thesis still needs earnings to back it up.
What to watch next?
Two threads will determine whether these targets are met: whether orbital data centres land on schedule, and whether terrestrial AI compute commercialises at the expected pace.
In plain terms = space and ground are two legs of the same race — whichever delivers real revenue first will move the stock first.
Earnings and capacity updates over the next few quarters are the most direct checkpoints.
市场有风险，内容仅供研究参考，不构成投资建议。