SpaceX Beat Expectations in Its First Earnings Report Since Going Public. Here's Why the Market Still Isn't Thrilled

SpaceX reported Q2 revenue of $7.8B (up 92% YoY), beating the $6.9B analyst expectation, while reducing its loss to -$0.09 vs a forecast -$0.26. However, investors focused on capex of $18.4B for the quarter (vs $2.8B a year ago), with most of the spend tied to AI, sparking a double-digit stock drop in early trading. Despite the operational beat, the stock’s ~$1.7T valuation leaves limited “margin of safety,” and the article suggests downside risk could push shares below $100.

Analysis

This is not a revenue problem; it is a capital efficiency problem. When a growth story starts requiring multiples of prior capex to preserve its runway, the market stops rewarding top-line acceleration and starts discounting future dilution, lower free-cash-flow conversion, and a longer payback period. That typically hits the highest-duration names first, then spills into adjacent AI spenders that also trade on "invest now, monetize later" narratives.

The second-order winner is the AI hardware layer, but only if spend is genuinely external procurement rather than internal experimentation. If the capex is going into GPUs, networking, and datacenter build-out, that is supportive for NVDA and the broader semiconductor supply chain over the next 1-3 quarters; if it is mostly internalized software/model training, the benefit is far less direct. The loser set is broader: unprofitable AI applications and private-market analogs will see multiple compression as investors demand evidence of unit economics, not just growth.

Near term, the key risk is that this becomes a crowded de-risking event for any name with expanding capex and no near-term FCF. Over 6-18 months, the thesis reverses only if utilization ramps enough to pull capex as a percentage of revenue materially lower; otherwise, the market will keep applying a higher discount rate and a lower terminal margin assumption. The contrarian read is that the market may be overreacting to front-loaded infrastructure spend and underestimating that some of this capex is pre-positioning for a larger connectivity moat, but that only matters if management can show rising gross profit per incremental dollar spent within the next two prints.

Key Decisions for Investors

• If SPCX is liquid/borrowable, fade relief rallies with a short or put spread into the next earnings catalyst; thesis breaks only if management cuts capex guidance or shows a sharp step-up in free-cash-flow conversion.
• Pair trade: long NVDA / short SPCX for 1-3 months. Rationale: capture AI infrastructure demand while avoiding the balance-sheet and execution risk of a capex-heavy integrator; this works only if capex is external hardware spend rather than internal R&D.
• Avoid chasing other high-multiple AI spenders until the market proves it will reward spend with near-term monetization; watch for multiple compression in any software AI proxy if SPCX keeps trading down on capex optics.
• Set an alert on SPCX for any quarter where capex falls below revenue growth for the first time; that would be the inflection point for a re-rating and would invalidate the short thesis.
• For event-driven traders, use the current dislocation to wait for a 5-10% intraday bounce before initiating bearish exposure; the risk/reward is better once post-earnings forced selling and retail capitulation are absorbed.