Samsung (SSNLF.US) plans to sign multi-year contracts to alleviate concerns over storage chip shortages

CryptoTimes App has learned that South Korea’s memory chip giant Samsung Electronics (SSNLF.US) is considering shifting toward signing multi-year memory chip contracts, with terms far exceeding the usual levels, which could help stabilize supply and ease concerns over shortages of this critical component.

Samsung Electronics Co-CEO Young Hyun Jun told shareholders at the annual general meeting that the company is considering extending current quarterly or annual contracts to a maximum of three to five years. He expects demand for AI storage chips to continue surging through 2026.

Samsung Electronics’ stock price temporarily soared 6.5% on the Korean exchange, marking its third consecutive day of gains, and boosting other group stocks. Samsung C&T Corp. and Samsung Life Insurance Co. each rose by as much as 8.9% and 13%, respectively.

For decades, the memory chip industry has experienced cyclical patterns of “boom—crash—recovery,” but now, the AI spending boom is pushing the industry into a new rhythm. Industry executives say AI has structurally broken the old cycle, with no signs of prices declining.

According to reports, South Korea’s memory chip giant SK Hynix states that the entire memory chip industry is undergoing a structural transformation. A company spokesperson said in a statement, “Our customers, including large-scale data center operators, are increasingly inclined to sign long-term contracts rather than the more common one-year agreements in the past.” Micron Technology (MU.US) also said that customers are now very willing to sign long-term supply agreements to lock in memory supply for the coming years.

By 2025, as global AI infrastructure construction enters a boom period, the memory industry is entering a “super cycle.” The core logic is that, on one hand, AI servers demand far more storage capacity and bandwidth than ordinary servers. On the other hand, industry capacity is shifting toward high-end storage products (such as HBM), squeezing traditional storage product capacity and triggering a price increase across the entire storage industry.

Currently, AI infrastructure development continues to accelerate, with major tech companies expected to spend a staggering $650 billion in 2026, an approximately 80% increase from last year’s record levels. Industry research shows that data centers’ demand for DRAM accounted for about 50% of global consumption in 2025, up from only 32% five years ago. This proportion is expected to continue rising. By 2030, AI servers are projected to account for over 60% of global storage demand.

AI demand is triggering a historic shortage of memory chips. Meeting the exponential growth in chip demand will be costly and may even be impossible to fully achieve. Although Micron, SK Hynix, Samsung Electronics, and others are expanding capacity through new or upgraded manufacturing facilities and advanced packaging plants, such projects often take years and billions of dollars to produce significant output.

For now, companies involved in data center construction are still securing the necessary memory chips to continue expanding. Their revenues and profits are growing alongside memory chip manufacturers benefiting from surging demand. However, for consumer electronics companies, this supply crunch could mean higher product prices, lower profit margins, and slower product upgrades.

Executives from companies including Apple (AAPL.US), Google (GOOGL.US), and Tesla (TSLA.US) are already discussing how the memory chip shortage could impact profitability and even the timeline of AI development. Google DeepMind head Demis Hassabis called it an industry “bottleneck.” During Tesla’s January earnings call, CEO Elon Musk even proposed the idea of producing their own storage chips.

SK Group Chairman Chey Tae-won stated on Monday that, as AI-driven demand continues to outpace supply, the global memory chip shortage could persist until 2030. Explaining the cause of the shortage, Chey Tae-won said, “AI actually requires a large amount of HBM, and once we start manufacturing HBM… we will need to consume a huge amount of memory chips,” “Therefore, it will take some time to increase more memory chip capacity, at least four to five years. The current shortage could last until 2030, and we expect the shortage rate to exceed 20%.”

He also mentioned that SK Hynix will strive to develop strategies to stabilize DRAM prices, “so I can’t announce it here directly, but I guess our CEO will announce a new plan on how to stabilize DRAM prices.”