Thinking of Buying SpaceX? Just Buy This Market-Thumping Stock Instead.

SpaceX, now roughly flat since its June IPO, is ramping AI compute to 10 GW by end-2027 (about 7.5M US homes), but remains expensive at ~64x sales and only expected to reach ~$0.09 EPS this year. The article argues Nvidia is the safer AI proxy: Nvidia reported ~106% quarterly revenue growth (about $50B added sales) and guided to 70%+ growth next year vs ~45% expected. It highlights an exclusive SpaceX partnership for Nvidia’s Vera CPU/Vera Rubin systems, implying Nvidia stands to benefit from SpaceX’s AI buildout while being priced at ~18x next year EPS.

Analysis

The clearest implication is not that SpaceX becomes a direct AI winner, but that its capital intensity makes it a leveraged buyer of compute while NVDA remains the toll collector. If the 10GW build is even partially real, the spend curve favors suppliers with the strongest product cadence and allocation power; that supports NVDA’s revenue visibility while squeezing weaker accelerators and custom-chip efforts at AMD and INTC.

The market is likely underestimating how fragile the private valuation is if AI timelines slip. SPCX is effectively a long-duration cash-flow story with two expensive bets layered on top of each other: launch/connectivity and AI infrastructure; any delay in monetizing the AI pivot should force multiple compression faster than it would for NVDA, whose earnings are already being recognized in public markets.

The contrarian angle is that exclusivity is not the same as structural demand. A single strategic relationship can look like a moat, but if SpaceX’s AI spend proves more aspirational than budgeted, the read-through to NVDA is incremental, not transformative; the real upside for NVDA is breadth of demand, not one customer. Near term, this is more of a sentiment-positive confirmation for NVDA than a fundamental rerating catalyst, while SPCX looks vulnerable to disappointments over the next 1-3 quarters.

Key Decisions for Investors

• Long NVDA / short SPCX as a relative-value pair over the next 1-3 months; thesis is that public-market earnings visibility and valuation support NVDA, while SPCX’s AI optionality is priced like execution is already de-risked. Falsify if SPCX shows audited capex commitments or backlog that credibly supports the 2027 compute target.
• Use NVDA call spreads rather than outright calls into the next earnings cycle; the article supports continued multiple resilience, but the better setup is still earnings-plus-guidance confirmation, not a re-rating from one strategic announcement.
• Avoid chasing SPCX at current levels; if accessible, treat any post-rally strength as an opportunity to fade until management provides hard budget, financing, and deployment milestones. Risk/reward is poor unless there is evidence of externalized AI revenue within 2-3 quarters.
• Watch AMD and INTC as secondary shorts on any sign that SpaceX/xAI is locking in Nvidia-only architecture. The trade works only if procurement decisions become durable; if SpaceX diversifies accelerators, the read-through disappears.