The Export Controls Worked. That Is Exactly Why They Are Failing.

On 5 August 2026, according to the Korean outlet Digital Daily, Apple asked ChangXin Memory Technologies for a price cut on LPDDR5X mobile DRAM and was refused. CXMT quoted terms at or above what Samsung and SK Hynix already charge.

You do not have to take the report on faith. On JD.com, a 64GB DDR5-5600 ECC RDIMM built on CXMT memory lists at ¥18,999. The identical module on Samsung or SK Hynix silicon lists at ¥18,595. Same capacity, same speed, same retailer, and the Chinese part is the expensive one.

Which inverts the entire policy debate. The Chinese supplier — the one seven senators want Apple to swear off, the one Commerce has drafted for the Entity List — did not undercut anybody. So Apple is not lobbying Washington to save money. It is lobbying for volume that exists nowhere else at any price.

And the reason CXMT can hold that line is the part nobody in the argument will say out loud: Washington built it.

Export controls have kept CXMT off EUV lithography. Denied those machines, it prints the same features by multi-patterning — repeated exposures with older 193nm DUV tools, each slightly offset, approximating in several passes what EUV does in one. Every pass costs time, chemicals, energy and yield. The arithmetic: CXMT needs roughly 30% more wafer starts than an EUV-equipped rival to ship the same memory. Morningstar flagged that same figure as a hard cost ceiling in July, during the IPO — before the Apple story existed, for unrelated reasons.

A producer carrying a structural 30% penalty on cost per bit cannot run a price war. It is not a strategic choice, and Chinese state money does not fix it, because the constraint is not capital. It is a machine that cannot be shipped.

The controls did precisely what they were designed to do. And that success is now the problem