SpaceX's Crash Didn't Start With Flight 13

SpaceX shares first fell below their $135 IPO price on July 7. The post-IPO quiet period had just expired, and Wall Street's first batch of analyst coverage landed with a spread so wide it was itself a verdict on the stock's central claim. Morgan Stanley set a $300 target, projecting $3.3 trillion in revenue by 2040. Raymond James went to $800. Morningstar issued a Sell at $62. MoffettNathanson set $131 ^[1].

sufficient material inputs will not exist — MoffettNathanson

The stock fell roughly 7% that day, closing about $11 below the IPO price. Nine days later, on July 16, four Raptor engines failed to ignite on Starship Flight 13, the vehicle aborted, and shares slid further — to $131.11, then $124.30 the next day ^[2]. The sequence looked like cause and effect: an engine failure, a stock rout. But the break below IPO had already happened. Flight 13 did not trigger it; it only deepened a decline that was already underway. What was actually underway was a synchronized repricing of the entire AI infrastructure complex. The ledger is worth lining up. On June 3, Broadcom's cautious forward guidance triggered a $1.3 trillion single-session sell-off in AI stocks ^[3]. On June 26, OpenAI delayed its IPO, and global AI shares crashed again — South Korea's Kospi dropped 9%, Japan's Nikkei fell 4%, SoftBank plunged 12% ^[4]. On July 14, IBM fell 25% in a single day, its worst decline in 115 years, after CEO Arvind Krishna told investors the company had been caught flat-footed by a structural shift in enterprise spending toward AI hardware ^[5].

This quarter we faltered. — Arvind Krishna

The same week, the global semiconductor index fell 11%, nearing bear-market territory — SK Hynix dropped more than 15%, and the rout spread to Samsung, Intel, Micron, and ASML ^[6]. On July 17, the day SpaceX closed at $124, Apple overtook Nvidia as the world's most valuable company, marking a broad rotation out of AI infrastructure providers ^[7]. These are not separate events. They are one event — a sector-wide deleveraging — and SpaceX's 45% decline from its $225.64 peak is one row in the same ledger. The counterintuitive part is what did not happen. A crash of this magnitude, in a company whose launch-and-satellite business lost $5 billion on $18.7 billion in revenue last year, should have re-coupled the stock to something resembling launch economics ^[8]. It hasn't. At $124 a share, SpaceX is still valued at roughly $1.6 trillion. Even its closest satellite-connectivity competitor, AST SpaceMobile, has been repriced downward by 60% — and that company projects $1 billion in revenue by 2027 ^[8]. The entire sector is being marked down, but SpaceX is not being singled out for a return to fundamentals. If it were, the number would be far lower than $1.6 trillion, because the fundamentals do not support even that. The gap this column identified last week — Starlink's profits justifying roughly 4% of the IPO price, the rest a bet on orbital AI and Mars — has widened only modestly, to perhaps 7% to 9% at the current price. The AI premium has survived a 45% crash, an engine failure, and a break below the IPO price. It persists while Musk deepens rather than retreats from the AI infrastructure build-out: in May he bought APR Energy for $1 billion to deploy 59 mobile gas turbines at the Colossus II AI data center, and the Pentagon has argued against an environmental shutdown of those turbines on national security grounds because the military relies on the Grok AI models hosted there ^[9]. What Flight 13 did was narrower and more specific. Musk announced the Starmind project — orbital AI data centers, targeted for 2027 — tying Starship's technical progress directly to the AI infrastructure thesis ^[10]. Starmind depends entirely on Starship working. Four Raptor engines failing to ignite is a technical question mark on that specific premium. But the market had already been trimming it for weeks: by July 11, five days before the abort, shares were at $148, down 34% from peak, and the AI segment had reported a $2.5 billion first-quarter operating loss ^[11]. The engine failure confirmed a doubt the market had already priced in. The premium's persistence is not even SpaceX-specific anymore. Blue Origin is now raising $10 billion at a $130 billion valuation, and its pitch includes "Project Sunrise" — a proposed network of 51,600 satellites for orbital AI data centers, directly mirroring Starmind ^[12].

We are entering a period where opportunities in space are expanding rapidly. — Dave Limp

A pure launch company is now selling itself on the orbital AI premise, not on launch economics. The market has not re-coupled SpaceX to the launch business. It has barely begun to narrow a gap that may be too wide to close unless the orbital AI thesis actually works — and that thesis now belongs to the whole sector, not to SpaceX alone.