Driven by explosive growth in AI infrastructure construction demand, the global memory chip market is facing severe supply shortages. Leading technology companies like Dell Technologies (DELL) and HP (HPQ) have successively warned that persistent supply constraints are likely next year due to surging demand for AI-related chips. Market research firm Counterpoint Research predicts that memory module prices could rise by up to 50% by the second quarter of next year, reflecting the severity of the supply-demand imbalance.
As a core component of modern electronic devices, the tight supply of memory chips is rapidly spreading to multiple sectors including mobile phones, medical equipment, and automotive manufacturing. Driven by the AI wave, chip manufacturers are prioritizing production capacity for newer, higher-margin, high-bandwidth memory products, leading to a squeeze on the supply of general-purpose memory. Jeff Clarke, Chief Operating Officer of Dell Technologies, noted in a recent analyst meeting that the company has never experienced such rapid cost fluctuations, with key components like DRAM and NAND flash memory facing tightening supply. He admitted that although the company would mitigate pressure by adjusting its product mix, the impact of rising costs would ultimately be passed on to consumers.
Since ChatGPT ignited the generative AI boom, the global wave of data center construction has accelerated the strategic transformation of the memory chip industry. Giants like Samsung and SK Hynix, which control about 70% of the global DRAM chip market, are gradually shifting production capacity from traditional products like DDR4 to premium chips such as DDR5 and high-bandwidth memory. According to industry sources, these manufacturers plan to completely stop production of DDR4 memory chips by the end of 2025 to early 2026. This structural adjustment coincides with the upgrade cycles of traditional data centers and personal computers, coupled with better-than-expected mobile phone sales, collectively exacerbating the tight supply of non-high-bandwidth memory chips. Data from Morgan Stanley shows that tech giants’ investments in AI infrastructure this year will reach a scale of $400 billion, further reinforcing this trend.
Faced with the intensifying supply crisis, technology companies are exhibiting different coping strategies. Enrique Lores, CEO of HP, expects the second half of 2026 to be particularly challenging; the company is preparing to alleviate pressure by introducing new suppliers and optimizing product design. It is worth noting that memory costs now account for 15%-18% of the total cost of a typical PC. In contrast, Apple has provided a relatively optimistic assessment, as its advantageous position in the supply chain helps secure favorable supply terms.
