SpaceX is falling after posting first earnings report since IPO. What the biggest Wall Street firms are saying

SpaceX (SPCX) reported a Q2 loss of $0.09/share, better than the $0.26 loss expected, alongside revenue of $7.81B vs $6.93B consensus. Despite the beat, shares fell over 12% as AI spending surged to $23B+ in the first six months from $3.3B a year earlier, lifting capex. The company forecast >$100B in annual recurring revenue by year-end, and multiple analysts reiterated buy ratings with upside of ~88% (e.g., BofA $235 target), though near-term volatility is expected given elevated capex and AI/Starlink monetization questions.

Analysis

The market is treating the selloff as a funding problem, but the real question is whether each incremental unit of compute is being pre-sold at a return that compounds faster than the balance sheet expands. If the implied revenue-per-watt trajectory is even directionally correct, then the capex step-up is not dilutionary in an economic sense; it is a capacity bottleneck being financed ahead of demand. That framing supports a premium multiple, but only if backlog converts to cash rather than remaining narrative.

The clearest second-order beneficiary is NVDA: an exclusive accelerator relationship plus constrained power/buildout makes GPU supply the scarce input, not the output. A less obvious loser set is MSFT, GOOGL, and META if this evolves from a niche story into a credible alternative compute layer; the immediate impact is not lost revenue, but multiple pressure as investors question how durable hyperscaler pricing power is in frontier training and inference. If SPCX can undercut frontier compute economics, it could also pull spend away from the same AI budgets that have justified elevated capex across the megacap cohort.

Near term, volatility should stay high because the stock is being repriced on liquidity and lockup mechanics, while the fundamental debate will take months to resolve. The key falsifiers are delayed launch cadence, power shortages, or a capex ramp that is not matched by disclosed contracted revenue and collections. Over 6-18 months, the structural bull case depends on whether the company can translate pre-announced capacity into recurring revenue with operating leverage rather than perpetual reinvestment.

Key Decisions for Investors

• Buy SPCX only on post-lockup weakness, with a 3-6 month horizon into the next launch/contract disclosure; thesis works if management can reiterate YE26 $100B ARR and show cash conversion, and fails if backlog remains opaque.
• Run a relative-value long NVDA / short MSFT or GOOGL basket for 1-3 months; SPCX validates tighter GPU supply and should support NVDA pricing power, while hyperscaler multiples are more exposed if specialized compute gains credibility.
• If conviction on monetization is low, use SPCX call spreads instead of common stock into the next catalyst window; that caps downside if liquidity overhang persists while preserving upside if the revenue-per-watt narrative proves out.
• Watch META and GOOGL capex commentary over the next earnings cycle as a confirmatory signal; higher AI infrastructure spend would validate the scarcity thesis and reduce the odds that SPCX's growth is purely narrative-driven.