SpaceX's Terrafab Threatens Memory Oligopoly, Adding Risk to Lam Research's Equipment Demand Thesis

What happened

SpaceX has unveiled renderings of a 100-million-square-foot Terrafab facility, designed to break the memory oligopoly held by Micron, SK Hynix, and Samsung, posing a direct threat to the core customers of wafer fab equipment supplier Lam Research. If successful, this new capacity could disrupt memory pricing and investment cycles, reducing equipment demand from incumbent manufacturers that account for a significant portion of Lam's revenue. The announcement adds a new layer of uncertainty to Lam's already concentration-heavy outlook, which is highly dependent on sustained DRAM spending and faces ongoing China export-control risks. While details remain speculative and execution is uncertain, the sheer scale of the project signals a potential long-term structural shift in the memory market that equipment investors cannot ignore. This development reinforces the cautious WAIT rating, as it further clouds the visibility of a durable AI-led equipment upcycle.

Implication

The Terrafab announcement challenges the assumption that Lam's memory customers will maintain their current investment trajectories, as a new, well-funded entrant could compress margins and delay capacity plans. This risk is layered on top of Lam's already precarious dependence on DRAM-heavy growth and a China market where export restrictions are tightening. While the factory's impact is years away and its feasibility unproven, the market is likely to begin pricing in the uncertainty, keeping a lid on multiple expansion. Investors should monitor execution milestones and customer responses, as any shift in memory capex guidance from incumbents could be a catalyst. Until clarity improves, the elevated risk profile supports a patient approach and a lower entry point.

Thesis delta

SpaceX's Terrafab plan introduces a new, long-term disruption risk to the memory oligopoly that underpins Lam's equipment demand. This adds to the existing concerns of concentrated China exposure and narrow AI-led DRAM cycle, tilting the risk/reward further negative. The WAIT rating is strengthened, with an even lower probability of a durable upcycle if memory incumbents face a formidable new competitor.