SpaceX beat cleanly, and capex sank it anyway.
The same night's two biggest earnings prints (SpaceX and AMD) both beat cleanly and both sold off, and the shared mechanism is that investors have stopped rewarding an AI-capex beat on its own terms — they are pricing the capex line itself as a cost center until proven otherwise. SpaceX's AI segment tripled revenue but posted a $1.3B operating loss against $15.8B of AI-infrastructure capex specifically — a six-fold YoY jump — and the stock's intraday +9.4%-to-after-hours -8.6% swing shows the market reacting to the capex number, not the revenue beat. This directly validates the "zero-AI-value" framing already tracked on this thread (Morgan Stanley's 07-24 note argued a ~$100 share price implies zero value for the AI segment).
Scenarios 2
The market's capex-as-cost read hardens across the sector — more beat-and-raise prints get sold off on capex/margin lines rather than rewarded on revenue, becoming the standing rule for how AI-infrastructure names get priced this earnings season.
SpaceX and AMD are two independent, same-night instances of the identical pattern (beat cleanly, sell off on the capex/margin story) — two data points on the same night is a real signal, not noise, especially when the two companies don't compete directly.
Precedent — Oracle's equity-vs-credit divergence (tracked on this map since 07-09/07-17) already showed credit markets pricing distress from AI capex while equity still rewarded growth — SpaceX/AMD's same-night sell-off is the equity market itself starting to adopt the credit market's skepticism.
The AI segment's growth rate (nearly tripled YoY) proves out over the next 1-2 quarters into profitability or a clearer path to it, and the market's capex punishment this print is revealed as premature — a buying opportunity rather than a repricing.
Management guided to a $100B annualized revenue run-rate by end of 2026; if that materializes with the AI segment's loss narrowing, the capex will have been the necessary front-loaded cost of the growth the market wanted, not a red flag.
Context
First public earnings since SpaceX's IPO, so there is no same-company precedent for how the market treats its prints — the AMD comparison the same night is the closest available reference. The insider-lockup first tranche (up to 20% of ~911.5M restricted shares) follows 08-06, two trading days out, and will test whether this sell-off compounds into real supply pressure.
Threads 1
This is generated reasoning, not a sourced fact — the mechanism and scenarios above are the model's read on what this item could mean for capital flows, tagged with its own confidence rather than stated as settled. Back to Global Capital.