Morgan Stanley Reiterates SpaceX $300 Price Target, Starlink Contributing 85% of Profits
nashnova research
Morgan Stanley reiterated a $300 price target and Buy rating on SpaceX, implying ~112% upside; the core thesis is that Starlink delivered 85% of group operating profit in Q3, doubling year-on-year — but whether the target holds depends on the enterprise-AI compute ramp.
Why is Morgan Stanley calling for a double?
Analyst Adam Jonas's team reiterated a $300 target on SpaceX, implying ~112% upside from the $141.70 close.
SpaceX rose 3.16% on the day while the Dow fell 1.64%, the S&P 500 dropped 1.03%, and the Nasdaq slid 0.74%.
This means → the market is already voting with real money: even on a broad sell-off day, capital kept flowing into SpaceX.
What makes Starlink the backbone of the valuation story?
Starlink-led connectivity posted $2.367 billion in operating profit in Q3, up 99% year-on-year — roughly 85% of the group total.
Revenue comes from subscription fees paid by households, enterprises, and governments — recurring, not one-off contracts.
In plain terms = Starlink is no longer a cash-burning satellite project; it is a profit machine. However AI valuations swing, the subscription fees keep coming in.
How does the $300 target break down?
Morgan Stanley uses a sum-of-the-parts model: launch services $8/share, Starlink connectivity $111/share, X platform + Grok $11/share, enterprise AI compute $170/share.
The AI segment is already halved to account for execution risk, yet still represents 56.7% of the target.
This means → more than half the target rests on AI compute leasing — Starlink is the floor, AI is the ceiling.
How much can one Starlink launch earn?
At projected 2030 operating levels, one fully loaded Starship launch of V3 broadband satellites — assuming ~5% of total capacity is ultimately sellable — can support $300–400 million in annualized revenue.
Measured by incremental EBIT, the payback period is roughly 20 weeks.
In plain terms = one rocket pays for itself in under half a year — provided the added bandwidth reaches regions with paying demand.
How confident is Morgan Stanley on the AI compute ramp?
Morgan Stanley projects 8.3 GW of live compute capacity by end-2027, below Musk's 13 GW target; rising to 14 GW by end-2028.
Revenue per watt is expected to fall from $34 in Q1 2027 to $21 in FY2028 — scale goes up, unit price comes down.
Stripping out a one-time deferred-tax gain, SpaceX trades at roughly 17.5× 2028 P/E, slightly above the S&P 500's 16×; consensus is more bullish, implying about 11.2×.
This reflects a notably conservative stance from Morgan Stanley — yet even so, they still arrive at a double-your-money target.
What could derail the thesis?
Terafab wafer-fab plan: capital expenditure is enormous — $15 billion in 2027, cumulative $170 billion by 2030, $490 billion by 2035. Shareholders need a visible path from spend to utilization.
SpaceX–Tesla capital relationship: Morgan Stanley flags this as its single biggest portfolio concern — some investors see industrial logic in a potential combination, but uncertainty over timing, board independence, and dilution keeps them on the sidelines.
Peers remain bullish too: Goldman at $245, Citi raised to $280, Cantor at $345. The common question across all targets is whether AI compute can sustain adequate utilization as per-watt revenue declines — that is the make-or-break validation point for $300.
市场有风险,内容仅供研究参考,不构成投资建议。
