Micron (MU) plans to roughly double its HBM production capacity to about 100,000 wafers per month by year-end, ET reports

Micron (MU) plans to roughly double its HBM production capacity to about 100,000 wafers per month by year-end, ET reports.

Capacity announcements in high-bandwidth memory have become the recurring battleground of the AI buildout, and each of the three HBM producers has used staged expansion headlines to signal commitment to share in what has been a supply-constrained market. Historically, the sequence in memory is well worn: tight supply and elevated pricing, aggressive wafer additions from all suppliers, then the question of whether demand absorbs the new bit supply or the cycle reverts to the oversupply pattern that has defined commodity DRAM for decades. The relevant distinction here is HBM versus conventional memory: HBM carries long qualification cycles, advanced packaging bottlenecks, and contracted volumes, so wafer capacity does not translate directly into sellable supply, and yield rather than nameplate capacity has been the binding constraint in past ramps. For Micron specifically, the tell is how much of the new output is already committed under customer agreements versus speculative, since uncommitted capacity reads differently for pricing. The follow-ons are peer responses on their own HBM roadmaps, customer qualification milestones, and any commentary on bit supply growth versus demand in upcoming earnings calls.