Samsung is accelerating one of its most aggressive memory capacity buildouts in years, aiming to bring its Pyeongtaek Line 4 fab, or P4, into full operation by the end of 2026 as the AI server boom reshapes demand for high-bandwidth memory (HBM) and advanced DRAM.

As enterprise adoption of generative AI accelerates, a new phase of infrastructure demand is beginning to take shape. According to DIGITIMES' special report, Accelerating enterprise AI: Hardware advancements and compute architecture transformation, the industry is moving beyond the initial buildout of training capacity and into a stage defined by large-scale deployment—where inference workloads are emerging as the primary driver of compute growth.

Japanese auto parts supplier Denso said on April 27 that it is considering all options, including withdrawing its acquisition proposal for chipmaker Rohm, after failing to secure the company's support.