Cathie Wood Keeps Buying SpaceX Stock. Here's Why I Wouldn't Buy It Above a $1.8 Trillion Valuation.

Starlink delivered $4.3B in Q2 2026 revenue with a 38.6% operating margin and subscriber growth reportedly doubling YoY, but SpaceX posted negative free cash flow of about $25B in the first half of 2026. The article attributes the cash burn largely to heavy AI investment, with AI capex near $16B over one quarter as AI revenue reportedly rose from $818M to $2.6B after serving Anthropic. Despite Cathie Wood/Ark Invest buying on IPO day, the piece argues the valuation (notably “above $1.8T”) leaves limited margin of safety given AI’s longer-term return uncertainty.

Analysis

The key market mechanism is that the equity is behaving more like a capital-intensive infrastructure platform than a software compounder: the more aggressively it scales, the more the marginal dollar of growth is being absorbed by upfront investment. In that regime, the winners are the suppliers of the capital stack — especially high-end compute and datacenter infrastructure names like NVDA and the broader AI supply chain — while the loser is the equity holder who underwrites duration risk without clean visibility on payback. For competitors, the most important second-order effect is that a sustained service-quality lead in satellite connectivity can pressure smaller alternates and delay any rational pricing recovery in the space connectivity niche.

Near term, the stock is vulnerable to multiple compression rather than an outright fundamentals reset. The 1-3 month catalyst path is any update that confirms spend is still outrunning monetization, or that launch/next-generation deployment milestones slip again; either would push the market to assign a higher probability to permanent maintenance-like capex. Over 6-18 months, the structural question is whether cash burn normalizes as capacity scales, or whether AI/network investment becomes a quasi-opex treadmill. If the latter, the equity deserves a much lower terminal multiple and possibly a financing discount.

The contrarian view is that consensus may be underestimating the optionality embedded in a dominant network plus launch franchise: if the spend is front-loaded and utilization rises, today’s cash burn could still be value-creating. But the burden of proof is high, and the current setup looks more like an expensive call option on execution than an attractive compounder. The thesis would be falsified by evidence that capex intensity is peaking while segment margins and free cash flow trajectory improve over the next two quarters.

Key Decisions for Investors

• If SPCX liquidity is usable, fade strength with a 1-3 month put spread into the next quarterly update; the edge is in valuation compression, not a bearish call on the business model. Invalidated if management shows capex deceleration and FCF inflecting materially positive.
• Prefer a relative-value long NVDA / short SPCX basket on rallies: NVDA captures the infrastructure spend with far less balance-sheet risk, while SPCX is forced to defend a very high terminal-growth narrative. Cover the short if SPCX signals capex normalization or if NVDA demand commentary softens.
• Set an alert on SPCX for any disclosure of capex below-trend for two consecutive quarters or launch cadence improvement; that is the point where the 'treadmill' thesis weakens and the stock can re-rate sharply higher.
• For less direct exposure, monitor ASTS and IRDM as competitive barometers; if Starlink continues to expand without pricing erosion, those names remain structurally pressured, but a slowdown in subscriber growth would be the cleaner long setup than an outright SPCX short.