The rapid development of artificial intelligence (AI) has led technology companies to invest heavily in computing power. They now face a new challenge, however: a shortage of memory capacity, reports CNBC. That bottleneck is especially evident now that AI applications are shifting from model training to practical use, partly driven by the rise of agentic AI, which requires substantial memory for context and continuous learning.
Analysts at Morgan Stanley predict that this memory constraint will persist until 2027, which will lead to higher prices and favourable market conditions for companies involved in memory production and semiconductor manufacturing. They recommend investing in memory-related stocks, particularly those focused on DRAM (dynamic random-access memory) and EUV (extreme ultraviolet) technology.
The analysts cite Samsung, Micron and SK Hynix as key players in the DRAM market thanks to their pricing power and growth potential. They also suggest considering Winbond, a Taiwanese company positioned to benefit from rising demand for older memory types such as DDR4. Western Digital is another promising option, which could profit from higher prices for hard disk drives and enterprise NAND Flash memory, as AI workloads migrate to more cost-effective storage solutions.
In addition, Morgan Stanley identifies Disco, a Japanese manufacturer of chip-processing equipment, as a beneficiary of the growth in high-bandwidth memory (HBM), which is crucial for advanced AI applications. Applied Materials and ASM International are recommended for their role in supplying equipment for semiconductor production, particularly for the expansion of DRAM capacity. Finally, ASML, the only supplier of EUV technology, is expected to reap substantial gains from the increasing use of EUV in chip manufacturing.
Business AM
Ltd
