Shares of Space Exploration Technologies fell after hours Tuesday despite beating earnings expectations, as investors zeroed in on a sharp jump in capital expenditures tied to its artificial intelligence ambitions. SpaceX reported quarterly capex of $18.4 billion, up from $2.8 billion a year earlier, while management also signaled continued heavy spending ahead.
While the company’s results benefited from new data center contracts with AI customers such as Google and Anthropic, the magnitude and pace of investment helped drive the post-earnings pullback and sent attention across the semiconductor supply chain, particularly to memory makers.
Key takeaways
- Price move: SpaceX shares dropped 8% after hours following the earnings release.
- Catalyst: Capital expenditures surged to $18.4 billion from $2.8 billion in the year-ago quarter.
- Demand signal: CEO Elon Musk said AI memory demand is growing around 200% per year, far outpacing supply growth.
- Implication for investors: Musk’s remarks reinforce the view that memory pricing and the memory cycle could remain elevated.
What drove the move
SpaceX’s quarterly earnings beat expectations on both the top and bottom lines, supported by additional data center contracts from AI-focused customers including Google and Anthropic. The operational momentum appeared to validate the company’s growing role in supplying compute infrastructure for AI workloads.
However, the market’s focus shifted quickly to spending. The company’s capex rose to $18.4 billion in the quarter from $2.8 billion a year earlier. For investors, that increase matters because it signals both the scale of SpaceX’s planned infrastructure buildout and the potential pressure that elevated spending can place on near-term free cash flow.
In addition, management reiterated an aggressive growth target, saying it expects to reach $1 trillion in revenue by 2030. Hitting such an outcome typically requires sustained capital deployment—an element investors appeared to weigh heavily when pricing the post-earnings outlook.
What Elon Musk said about memory
Beyond the earnings figures, Elon Musk provided commentary that landed directly on the memory market. He said AI compute demand is rising much faster than memory supply, stating that memory output is increasing by around 20% per year while demand is increasing by about 200% per year, “maybe higher.”
Musk also framed memory as the limiting factor on growth for SpaceX and other AI infrastructure providers. His logic followed a straightforward supply-and-demand argument: when demand accelerates faster than supply, pricing pressure tends to increase, according to standard economics principles.
He also discussed SpaceX’s approach to compute architecture, attributing its strategy to Nvidia and saying the company would build exclusively on Nvidia because it views the Vera Rubin architecture as the best AI computer design.
Market reaction and second-order effects
Although the company’s results were strong in absolute terms, investors appeared reluctant to look past the capex ramp. The 8% after-hours decline suggests that the spending trajectory outweighed the immediate earnings beat as the dominant variable for near-term valuation.
At the same time, Musk’s comments about memory supply constraints have broader implications for semiconductor companies tied to AI servers. The report said much of SpaceX’s capex is directed toward chips, which links the company’s infrastructure expansion to demand for memory components used in AI compute systems.
For memory-related equities, that creates a potential tailwind. The article argued that if AI demand continues to outstrip memory output growth by a wide margin, memory pricing could remain upward pressured over the coming years, supporting the broader memory cycle.
Where Micron goes from here
Even with renewed strength in AI-linked stocks and the Nasdaq Composite hovering near all-time highs, Micron Technology remains down nearly 30% from its end-of-June peak, according to the article. That context matters because it suggests investors have already scaled back parts of the memory-cycle outlook after a recent sell-off.
The key question for the memory space is how long and how high the memory boom can last. While the market appears to be resetting expectations, Musk’s stated demand-growth gap—demand up roughly 200% per year versus supply output up about 20% per year—provides a concrete signal that the underlying imbalance may persist.
If that imbalance holds, it would support the investment case for memory providers as AI infrastructure buildouts intensify across hyperscalers and specialized compute operators. Conversely, any improvement in supply growth or a slowdown in AI demand would be central to how investors would adjust future expectations.
Investors will likely look next for additional guidance on capex intensity and for commentary on how quickly memory supply can respond to accelerating AI buildouts. Upcoming catalysts include more earnings updates from memory suppliers and AI hardware ecosystem players, alongside closely watched macro data and central bank signals that can influence the discount rates applied to high-growth technology spending.
