Samsung Electronics has reportedly begun mass production of 236-layer NAND flash at its Xi'an plant in China, marking a key step in its transition to higher-layer memory as demand from artificial intelligence applications accelerates.
The global DRAM industry is approaching a structural inflection point, as traditional scaling methods struggle to deliver the performance gains required by artificial intelligence workloads. With next-generation architectures such as 4F-squared (4F²) and 3D DRAM facing rising complexity and potential delays, manufacturers are being forced to reassess near-term roadmaps and rely more heavily on incremental and material-level improvements.
espite a surge in demand driven by generative artificial intelligence, the fundamental economics of the memory industry remain largely intact. While high-bandwidth memory (HBM) has created a premium segment, the broader market continues to operate on standardized, high-volume production rather than structural product differentiation.
Samsung Electronics' next-generation high-bandwidth memory (HBM4) program is entering a customer validation phase, according to NewDaily, as major technology firms reportedly begin on-site audits of its production lines.
SK Hynix CEO Kwak Noh-jung is expected to meet Scott Guthrie, executive vice president of Cloud and AI at Microsoft, in Seoul this week to discuss memory supply, according to Maeil Business Newspaper, citing industry sources.
ChangXin Memory Technologies (CXMT) more than doubled revenue to about US$8 billion in 2025, as surging demand from artificial intelligence (AI) and data centers drove up memory chip prices, according to Bloomberg. The growth provides a boost to the strategically important Chinese chipmaker ahead of a planned domestic IPO.
The global semiconductor industry is set to surpass the US$1 trillion revenue mark in 2026, driven by accelerating artificial intelligence demand, but the milestone comes with mounting structural pressures—particularly in memory—that are beginning to ripple across the broader electronics ecosystem.
Google has introduced TurboQuant, a compression algorithm that reduces large language model (LLM) memory usage by at least 6x while boosting performance, targeting one of AI's most persistent bottlenecks: memory. The breakthrough lowers inference costs and expands deployment across cloud and edge environments.
SK Hynix plans to build more than KRW100 trillion (approx. US$66 billion) in net cash to support long-term investment in artificial intelligence (AI) infrastructure, as surging demand reshapes the memory industry and drives a new round of capital spending.
