Only the Scaled SurviveNarrow moat

The combined weight of technology, capital, and the learning curve produces a single decisive fact: in memory, only the scaled survive. The barriers are so high that a maker must be enormous — spending tens of billions, operating at the leading edge, manufacturing at massive volume and high yield — simply to stay in the game, and anything less is slowly ground away. This is the mechanism that culled the field from dozens of makers to three, and it is the same mechanism that now protects the three: their scale is both the ticket to competing and the barrier that keeps the field from growing.

Scale in memory compounds. The largest makers spread the colossal fixed costs of R&D and fab construction over the most bits, achieving the lowest cost per bit; they can afford the continuous investment each node requires; and they have the volume to justify developing new products like HBM. A subscale maker cannot match these economics and is squeezed out in the next downturn, when the low-cost producers survive and the high-cost ones do not. Micron's scale as one of the three largest memory makers is thus not merely a size advantage but the precondition of its survival and the guarantor of the oligopoly's stability. The uncomfortable fact is that Micron is the smallest of the three — Samsung and SK Hynix are larger, with deeper resources — so within the oligopoly Micron is the scale laggard, and the same 'only the scaled survive' logic that protects it also means it must keep pace with two larger rivals or risk, over time, the subscale fate it has so far avoided. Scale is Micron's shield, but it holds the smallest of the three shields — Samsung and SK Hynix both out-scale it¹.