SpaceX’s CEO Elon Musk found memory chips to be the main constraint on artificial intelligence (AI) expansion, putting Micron Technology in the spotlight.
While talking about SpaceX’s Q2 2026 earnings, he claimed that the “limiting factor currently is memory,” indicating a massive supply-demand imbalance in the semiconductor industry.
Musk suggested that the memory supply grows around 20% per annum, describing this rate as “fantastically fast” for a mature industry. But AI-driven demand is increasing at 200% per year.
“So, if you’ve got demand increasing much faster than supply, then economics 101 would suggest that the price increases. It does not decrease,” he said.
These statements are especially important for Micron Technology (NASDAQ: MU), which is the sole HBM producer in the United States and among only three manufacturers of high-bandwidth memory globally. The share price of Micron has rallied despite challenging market conditions likely to persist past 2027.
Musk also dismissed power supply as a constraint, suggesting that SpaceX aims to have “far more power, cooling and electrical equipment than we have GPUs.” Discussing GPU availability from Nvidia, he noted the company would receive only “a very small percentage” of their chips next year.
Additionally, Musk ruled out power as an issue, saying that SpaceX wants to be “far.
Analysts have forecasted that AI agents, which will undertake actions independently, would utilize around 120 quadrillion tokens a month by 2030—a figure 24 times the current rate. Such increased activity by AIs will require massive memory needs for functions including retaining context, data access, and program execution.
According to Bank of America, AI-related memory requirements could raise Micron’s earnings per share to $236 in its fiscal year 2030, much higher than consensus forecasts.
