The Memory OligopolyNarrow moat

The foundation of Micron's moat — the thing that turned memory from a capital-destroying commodity into a business worth owning — is the consolidation of the DRAM industry into an oligopoly of three. This is the single most important fact about the company, and it is why Micron has a moat to speak of at all. For decades, memory was a byword for value destruction: dozens of makers, in Japan, Korea, Taiwan, the United States, and Europe, poured capital into identical products, competed on price alone, and blew themselves up in wave after wave of overcapacity, driving prices below cost and bankrupting all but the toughest. The survivors of that decades-long war of attrition are just three: Micron, Samsung, and SK Hynix, who together took 89.7% of DRAM industry revenue in the first quarter of 2026.¹

Consolidation changed the economics. Three players — rather than thirty — can behave rationally: scarred by the memory of ruinous price wars, they compete primarily on technology and cost discipline rather than reckless capacity expansion, and they are collectively more restrained about flooding the market in a downturn. This is not collusion, which would be illegal and is not alleged; it is the natural behavior of a small number of experienced survivors who understand that suicidal competition benefits no one. The result is a commodity market that, uniquely in memory's history, can be profitable across a full cycle rather than only at its fleeting peaks — and that gives the survivors real, if cyclical, pricing power in the up-cycles.

The oligopoly is protected by the same barriers that created it: the crushing capital cost and decades-deep technological know-how required to make leading-edge memory mean the field cannot easily grow beyond three. A new entrant would need to spend tens of billions, climb a brutal learning curve, and survive years of losses against three scaled incumbents who would happily price a newcomer into oblivion — which is precisely why, for all the strategic importance of memory, no credible fourth Western or Korean player has emerged in decades. The structure is stable and self-reinforcing.

Two qualifications keep the moat narrow rather than wide. First, three is still enough for competition: the oligopoly's discipline is a behavioral choice, not a structural guarantee, and it has broken down before into painful price wars when one player grabbed for share or a demand shock hit — memory still busts, hard, as fiscal 2023's multibillion-dollar loss attests. Second, NAND flash — the other half of Micron's business — is a tougher, more crowded market of five or six players with weaker economics, so the clean three-player structure applies only to DRAM. The oligopoly is a real and durable moat that has genuinely improved memory's economics. But it is a moat around a commodity, and its pricing power is cyclical and behavioral rather than structural and permanent — the difference between a narrow moat and a wide one.