Apple lobbies US for clearance to buy chips from blacklisted Chinese supplier CXMT

<cite index="51-3,52-2">Apple is lobbying multiple federal agencies in the Trump administration for approval to purchase memory chips from ChangXin Memory Technologies (CXMT), a Chinese chipmaker placed on the Pentagon's 1260H list due to alleged ties to the Chinese government and military.</cite> <cite index="52-4,54-2">Apple has approached the Commerce Department and is seeking guarantees from the administration, citing volatile memory chip prices and its decision to raise hardware prices after being unable to absorb rising component costs.</cite>

<cite index="51-4">CXMT competes in mainstream DRAM, while Micron's investment dollars are increasingly directed toward high-margin HBM products where demand continues to exceed supply.</cite> <cite index="59-2">CXMT reported revenue growth of over 700% year-over-year in Q1 2026 and is preparing for an IPO in Shanghai worth roughly $4 billion.</cite>

<cite index="51-5,57-4">Apple previously attempted something similar in 2022 when it considered sourcing memory from another blacklisted Chinese manufacturer, YMTC, and received congressional warnings. While the 1260H list carries few immediate legal repercussions, both the Pentagon and Congress have shown skepticism toward expanded US-China memory trade.</cite> <cite index="55-3,55-5">The lobbying campaign comes after Trump and Xi Jinping met in Beijing last month, and it remains unclear if Apple would receive administration guarantees that CXMT won't later be added to the more restrictive Entity List.</cite>

For infrastructure architects and product teams, this signals that memory commodity pricing power has structurally shifted: even Apple—historically shielded from cost inflation through massive scale and supplier bargaining power—is now forced to pass costs to customers or seek geopolitical carve-outs. If Apple succeeds, it could shift memory market dynamics; if it fails, the supply constraint will remain a multi-year headwind for consumer hardware makers and may drive faster adoption of more efficient edge inference models.