Beyond Rockets: SpaceX's Real Growth Story Is Its AI Infrastructure Business

SpaceX says AI represents a $26.5 trillion opportunity within its $28.5 trillion total addressable market, with 2025 AI capex reaching $12.7 billion. It has already signed long-dated compute-rental deals with Anthropic, Alphabet, and Reflection AI that could generate more than $76 billion in revenue if fully executed. The article is constructive on SpaceX’s AI monetization potential, though it also highlights heavy capex and execution risk.

Analysis

The market should treat this less as a pure “AI demand” story and more as a power, network, and financing arbitrage. The near-term winners are not the moonshot space plays but the suppliers of high-density power, cooling, networking, and launch logistics that can monetize the capex wave before orbital economics are proven. If these colocated compute contracts are real and sticky, they create a backdoor validation of private AI infrastructure valuation multiples for firms that can underwrite large prepayments and tolerate utilization volatility.

The second-order effect is that this looks more like a balance-sheet stress test than a demand forecast. AI capex at this scale implies a long payback period, so the key variable is not revenue booked but whether compute utilization stays high enough to offset depreciation and financing costs once customer exit rights and pricing resets begin to matter. That creates a classic late-cycle risk: the headline TAM narrative can stay intact while unit economics quietly deteriorate if capacity growth outruns model-training demand.

For public comps, Alphabet is the cleanest way to express optionality because it benefits from both demand for cloud/AI infrastructure and potential launch/partnering leverage without having to justify the same level of speculative capex. Planet Labs is the higher-beta derivative if the orbital-data-center angle gains credibility, but it is a pure sentiment trade until a real payload pipeline appears. The contrarian read is that the market may be underpricing how much of this opportunity will be captured by existing hyperscalers and launch providers rather than by the company originating the thesis.

Key Decisions for Investors

• Long GOOGL vs. short a basket of speculative private-market AI infrastructure proxies for 3-6 months; thesis is that hyperscalers capture the first monetization layer while the more aggressive orbital narrative remains optionality with execution risk.
• Buy PL on pullbacks for a tactical 1-3 month trade only; use it as a high-beta satellite-data/space optionality expression, but size small because any delay in orbital commercialization can unwind the multiple quickly.
• Avoid chasing the space-linked AI narrative outright; prefer selling upside in names with capex-heavy AI stories via covered calls or call spreads, since utilization and depreciation mismatch is the most likely next-quarter disappointment.
• If GS equity remains strong, consider it as a secondary beneficiary through financing/advisory activity tied to large private infrastructure raises; 6-12 month horizon, moderate reward, low direct operational risk.
• Watch NVDA on any confirmed acceleration in data-center power deployments: a sustained capex step-up would be supportive, but the risk/reward is asymmetric only if orders translate into actual installed base growth within 2-4 quarters.