SK Hynix is in talks with Intel to make memory chips in the US, Reuters reports

SK Hynix is reportedly negotiating a potential deal to manufacture memory chips in the United States for the first time, potentially through leasing part of Intel's Ohio fabrication site. The arrangement could expand U.S. memory-chip capacity and deepen supply-chain localization, while providing Intel an avenue to utilize its Ohio manufacturing infrastructure. Talks remain preliminary and no financial terms or agreement timeline were disclosed.

Analysis

For INTC, a third-party memory tenant would be less about near-term wafer revenue than validation that its Ohio footprint can become an asset-light foundry/industrial-park platform. A lease could improve fixed-cost absorption and support the credibility of Intel Foundry’s external-customer pipeline, but the earnings impact is likely immaterial until contract terms establish tenant-funded tooling, utility commitments, and duration. The market should not award a material multiple re-rating on negotiations alone; the relevant 1-3 month catalyst is disclosure of binding economics alongside construction and subsidy milestones.

SK Hynix gains strategic optionality if a U.S. manufacturing presence improves qualification with hyperscale customers seeking geographically diversified AI-memory supply. The more important second-order implication is pressure on MU: domestic capacity has been a differentiator in securing U.S.-based AI and defense-adjacent demand, and a credible SK Hynix U.S. route could narrow that advantage over 6-18 months. Conversely, memory process integration is difficult and any arrangement lacking advanced-node/HBM relevance may be primarily political optics rather than incremental supply.

Consensus may overstate this as a clean win for Intel Foundry. Leasing shell capacity does not demonstrate that INTC can profitably manufacture SK Hynix’s proprietary DRAM/HBM process, and a tenant arrangement could highlight the gap between real-estate monetization and foundry execution. Falsifiers are a definitive agreement with minimum-payment guarantees and tenant capex, versus delays, a limited logistics/packaging scope, or any further reduction in INTC’s Ohio build-out commitments.

Key Decisions for Investors

• Treat INTC as a catalyst watch rather than a standalone long: add only after a binding agreement specifies multiyear minimum lease revenue and tenant-funded equipment; absent those terms, avoid chasing a headline-driven move. The key downside trigger is another Ohio schedule/capex revision.
• For a 6-18 month relative-value expression, monitor long MU / short SKHY only if the proposed U.S. activity is limited to leasing, back-end operations, or non-HBM production. This preserves MU’s domestic supply premium while avoiding a directional memory-cycle bet; exit if SK Hynix confirms U.S. advanced HBM-capable capacity or major hyperscaler volume commitments.
• If INTC rallies materially on reports without contract disclosure, consider a tactical short against SMH over a days-to-weeks horizon. Risk is a binding deal bundled with meaningful government support or a broader Intel Foundry customer announcement, which could justify a sharper multiple expansion.
• Track MU commentary on domestic AI-memory qualification, SK Hynix U.S. customer commitments, and Ohio subsidy disbursements during the next earnings cycle; these are more decision-useful than the initial negotiation report.