Lutnick Presses Samsung and SK Hynix to Build US Memory Fabs: What It Means for the Memory Cycle
Commerce Secretary Howard Lutnick used a concrete-pouring ceremony at Micron’s fab site in Clay, New York, to tell Samsung Electronics and SK hynix that they will build memory fabs in the United States. Not by regulation. By envy. Micron is leading, he said, so the Koreans will feel jealous and ultimately have no choice but to follow.
The remarks landed one day before SK hynix listed American Depositary Receipts on the Nasdaq, and ten days after Samsung and SK hynix announced a combined ₩800 trillion investment program in Korea’s Honam region. The sequencing is the story. Washington watched two companies that control roughly 60% of the global DRAM market commit a generational capex cycle to their home country, and responded by asking them to commit it somewhere else.
The Gap Lutnick Is Actually Pointing At
The Korean firms are not absent from the United States. Samsung is building a foundry complex in Taylor, Texas. SK hynix is putting $3.87 billion into an advanced packaging plant and an AI-memory R&D center in Indiana. Both are real, both are under construction, and neither is what Lutnick is asking for.
Packaging is the back end. Foundry is logic. What Washington wants is the front end of the memory business — DRAM and NAND wafer fabrication, the process nodes where the actual margin lives and where the actual know-how is embedded. HBM is stacked DRAM; the stacking is the visible part, but the value sits in the base die and the DRAM process underneath it. Moving packaging to Indiana moves the least defensible part of the stack. That is precisely why the Koreans were willing to do it.
Micron, meanwhile, has raised its US commitment to roughly $250 billion through 2035, with a stated target of producing 40% of its DRAM output domestically. That is the benchmark Lutnick is holding up, and it is a genuine one — Micron is the only major memory maker whose interests align cleanly with the administration’s manufacturing-share objective. President Trump has said he wants half of global chip capacity made in the United States. Current US share is somewhere north of 10%. The arithmetic does not close without Korean DRAM.
Why the Koreans Will Resist
Three reasons, in ascending order of durability.
The cost argument is the weakest one. US construction and labor run materially above Korean equivalents, but subsidies and tariff exemptions can be tuned until the spread closes. This is a negotiation, not an obstacle.
The technology-transfer argument is stronger. Memory is a process business. The recipes, the yield learning, the tool tuning — these do not travel as documents, they travel as people. Standing up a leading-edge DRAM fab in Arizona or New York means seeding an American workforce with Korean process expertise, and that workforce eventually goes somewhere. Samsung learned this lesson from the other side of the table decades ago.
The legal argument is the binding one. Korea’s national core technology regime restricts the offshoring of designated strategic technologies, and advanced memory — HBM included — sits inside it. Even if Samsung and SK hynix wanted to move DRAM fabrication to the United States tomorrow, Seoul holds a veto. That converts a corporate negotiation into a state-to-state one, and Korea has no reason to give away its single most important industrial asset for free.
The Part the Market Is Reading Backwards
Memory equities have been trading the shortage. DRAM and NAND pricing is running hot, HBM is cannibalizing conventional DRAM wafer starts, and the supply-demand imbalance has been the entire bull case. Political pressure to add capacity is being narrated as validation — governments are fighting over your product, therefore your product is scarce, therefore buy.
That reading is inverted on any horizon longer than a quarter.
What Lutnick is attempting is a coordinated, subsidized, politically-forced acceleration of global memory capacity. Micron at $250 billion. Samsung and SK hynix at a reported $880 billion combined across their global plans. Korea’s own Honam buildout. If these ramps land unsynchronized, the cycle extends. If they land together — and subsidy-driven capex has a strong tendency to land together, because everyone is responding to the same political signal at the same time — then the shortage resolves into a glut, and memory does what memory has always done at the top of a cycle.
Fabs take years. Output requires customer qualification after that. So the near-term shortage is intact, and nothing in this news changes 2026 pricing. But the terminal value of the memory supercycle thesis is a function of how much capacity gets committed while prices are high, and Washington is now actively lobbying to increase that number. Political pressure that pulls capex forward is bearish for the memory cycle, whatever it does for the memory headline.
Who It Is Good For
The unambiguous beneficiaries are the toolmakers. Applied Materials, Lam Research, KLA, ASML, and the hybrid-bonding equipment names do not care whether a DRAM wafer is fabricated in Pyeongtaek, Clay, or Yongin. They care about the total number of wafer starts installed. A trade dispute in which three governments compete to subsidize memory fabs is, from the equipment perspective, a bidding war for their order book.
The ambiguous beneficiary is Micron. Lutnick openly acknowledged that Sanjay Mehrotra may not be thrilled about inviting rivals onto US soil, and he is right to acknowledge it. Micron’s structural advantage in this administration is that it is the only American memory maker. Diluting that with Korean fabs in Texas and Indiana trades a policy moat for a supply-chain talking point. Micron gets the subsidies either way; it does not get the scarcity either way.
The party with the least to gain is SK hynix, which just raised roughly $26 billion in New York and now finds that the price of capital markets access is a conversation about where the fabs go. The ADR listing was supposed to fund Korean capacity. Lutnick has made it a down payment on American capacity instead.
That is the trade being negotiated, and neither side has said so out loud yet.