Loading market event…
Loading market event…
MajorMarket eventTue, Oct 272 min read
Long-term orders strengthen memory makers' pricing power while pointing to persistent cost pressure for AI infrastructure buyers.
SK hynix reported record third-quarter results on October 27, 2026: revenue of KRW 104.62 trillion and operating profit of KRW 83.33 trillion, an 80% operating margin. The company said AI-memory demand lifted key product prices, while sales of higher-value server products expanded.[1]
Sales of HBM4 also expanded. This latest generation of high-bandwidth memory stacks memory chips to feed data rapidly to AI processors. Long-term agreements now cover more than 50% of the company's 2027–2028 production capacity; management expects supply shortages to persist through 2030, according to reporting.[1][2]
The company also plans approximately KRW 100 trillion in shareholder returns during 2026, including distributing most of the gains from its Kioxia investment.[1]
AI customers committing to long-term purchases make production easier to plan and future revenue more visible for SK hynix. Making HBM uses far more factory capacity than conventional DRAM—the memory that temporarily holds data in use—so AI demand can also squeeze ordinary memory supply and support pricing for rivals such as Micron.[1][2]
Persistent shortages would encourage memory makers to expand production, creating demand for chipmaking equipment; equipment shares have already rallied on that prospect. For AI-server and cloud-service buyers, the same chain means securing supplies earlier and potentially paying more for memory. Shortages through 2030 remain management's forecast: future demand and new capacity will determine how long suppliers retain that pricing advantage.[3][2]
Rivals such as Micron stand to benefit from tighter DRAM supply and stronger pricing. Micron shares rose following SK hynix's results, though each producer's profit gains still depend on its product mix and available output.[2][4]
The shortage outlook strengthens the case for memory makers to build factories, creating potential orders for equipment suppliers. Lam Research, Applied Materials and KLA shares rallied on that prospect; orders still depend on customers approving spending.[3]
Server builders and cloud providers may need to contract earlier to secure memory. If supply keeps trailing demand, higher purchasing costs could raise expansion budgets or squeeze profits where buyers cannot pass those costs to customers.[1][2]