⚠ Know-How Can Leak, Especially to ChinaModerate threat

The learning curve is the most durable barrier precisely because it cannot be bought — but for the same reason it is the one most vulnerable to being slowly, deliberately acquired, and China's memory ambitions make that threat concrete. Because the know-how lives in engineers and organizational practice rather than in patents or machines, it can walk out the door: experienced memory engineers can be recruited away with enormous pay packages, manufacturing methods can be learned through partnerships or reverse-engineering, and a state that has decided memory is a strategic necessity has every incentive — and vast resources — to acquire the expertise by whatever means it can. China's emerging DRAM and NAND makers are pursuing exactly this, backed by state money and a national mandate.

The reassurance is that a decades-deep, organization-wide learning curve is genuinely hard to transfer — no single engineer, or even a group, carries enough of the distributed, tacit knowledge to recreate a world-class memory operation elsewhere, and the incumbents guard their know-how and keep advancing, so a follower is chasing a moving target. Export controls on advanced tools also slow China's climb. But over a long horizon, sustained poaching, determined national effort, and the sheer patience of a state indifferent to short-term returns can narrow the experience gap that protects Micron — not overnight, and not completely, but enough, eventually, to introduce a fourth, undisciplined player into the oligopoly. The learning curve is Micron's deepest structural protection, but it is a human barrier in a world where a powerful state is patiently, deliberately working to learn what it protects — the one form of the technology moat that determined subsidy and time can erode — which is precisely China's CXMT-and-YMTC strategy¹.