SpaceX enters September as a study in operational contrasts. The company’s rocket division is chalking up launch milestones at a record pace, while its crewed spaceflight program has hit an unexpected snag. Yet the real action — and the source of both investor enthusiasm and internal friction — sits firmly in the AI data-center business, where a management shakeup and a looming GPU delivery deadline are testing whether the company can execute on its most ambitious non-space bet yet.
A Leak Grounds Crew-13, But the Launch Cadence Rolls On
Technicians flagged an oxidizer leak in the Dragon capsule’s propulsion system on September 1 during pre-launch checks, forcing a postponement of the Crew-13 mission to the International Space Station that had been penciled in for September 12. NASA Administrator Jared Isaacman said on August 31 that repairs could take anywhere from two to four weeks, pushing a realistic launch window to late September or early October. The four astronauts slated for the flight — NASA’s Jessica Watkins and Luke Delaney, Canada’s Joshua Kutryk, and Russia’s Sergey Teteryatnikov — will have to wait.
The delay is a visible blemish on SpaceX’s human-spaceflight record, a program that carries outsized symbolic weight with NASA and the public even if it contributes little to the bottom line. But the broader launch operation has hardly blinked. A Falcon 9 lifted off from Vandenberg Space Force Base on September 1 carrying another batch of Starlink satellites — the company’s 104th orbital launch of the year and its 79th Starlink mission of 2026. Days earlier, on August 29, a Falcon Heavy had successfully deployed NASA’s Roman space telescope on a mission valued at $4.3 billion. Early September also brings the 14th Starship test flight, the third for the V3 generation of the vehicle.
A Leadership Shuffle in the Data-Center Division
The more consequential story, however, is unfolding well below orbit. Following engineering concerns and reliability problems at AI data-center sites in Tennessee and Mississippi, SpaceX has reorganized the leadership of its AI infrastructure unit, pulling in managers from its more mature rocket and Starlink operations. Wesley Salandro, formerly vice president of Falcon rocket and Dragon capsule production, and Logan McConnell, previously in charge of launch operations at Starbase, are moving into top roles in AI infrastructure.
The reshuffle comes with notable departures. Jake Palmer, Zach Wells, and Pablo Mendoza have left the company, while Brent Mayo has decamped to OpenAI and Liz Balke to Anthropic. The poaching of talent by direct AI rivals underscores just how competitive the hiring market has become in this corner of the industry — and how much SpaceX’s AI ambitions now depend on shoring up operational discipline in a division that has grown at breakneck speed.
The Numbers Behind the AI Pivot
The scale of that growth is hard to overstate. SpaceX reported second-quarter 2026 revenue of $7.8 billion, up 92% year over year, with AI revenue alone surging roughly 250%. The SpaceXAI unit had 1.4 gigawatts of computing capacity online in June and generated $2.6 billion in AI revenue for the quarter, a dramatic jump from $737 million in the year-ago period. The company has plowed about $15.83 billion into AI infrastructure.
Should investors sell immediately? Or is it worth buying SpaceX?
Yet the target of more than 2 gigawatts of capacity by year-end is looking increasingly shaky. Planned expansions beyond the Memphis hub, including a buildout in Texas, have reportedly slipped. And a hard deadline looms on September 30: SpaceX must have roughly 110,000 Nvidia GPUs in place under a monthly compute-capacity deal with Google valued at $920 million. Notably, that same date is currently earmarked for the Starship Flight 14 test — a scheduling coincidence that leaves little room for slippage on either front.
When complex systems face operational strain, the same principle applies on Earth as in orbit: rigorous risk management separates smooth execution from costly failure. For workplace leaders, a free toolkit with 41 ready-to-use templates and checklists can help document hazards and keep operations running safely. Download the free Risk Assessment Toolkit
Analysts See Upside, Market Remains Wary
Wall Street, for now, is giving SpaceX the benefit of the doubt. Oppenheimer analyst Tim Horan raised his price target on the stock from $250 to $280 on September 2, maintaining an “Outperform” rating. Horan cited the company’s competitive edge in building infrastructure and compute capacity in a market defined by shortages, pointing to strategic developments including the Grok model and the Cursor acquisition.
The market’s verdict is more mixed. The stock closed Friday at €127.36, down 1.1% on the day, though it has climbed 4.2% over the past week and a robust 36% over the last 30 days. That recovery leaves the shares trading roughly 7.7% above their 50-day average of €118.31 — but still about a third below the 52-week high of €194.46 touched in June. With annualized volatility running at 86%, investors are clearly wrestling with how to weigh the steady hum of the launch business against the promise and the growing pains of the AI bet.
The coming weeks will test whether SpaceX can thread the needle: getting Crew-13 off the ground, hitting the Google GPU milestone, and restoring confidence in its data-center operations — all while keeping Starship’s test campaign on track. The management shuffle suggests the company knows where its weak spot is. Whether the fix arrives before the September 30 deadline is another matter entirely.
Ad
SpaceX Stock: Buy or Sell?! New SpaceX Analysis from September 5 delivers the answer:
The latest SpaceX figures speak for themselves: Urgent action needed for SpaceX investors. Is it worth buying or should you sell? Find out what to do now in the current free analysis from September 5.
SpaceX: Buy or sell? Read more here...
