Investing.com– Memory semiconductor stocks have fallen sharply, but UBS analysts argue that concerns over future supply growth and competition may be overstated, leaving the sector attractively valued amid strong demand from artificial intelligence infrastructure.
Memory stocks have dropped 19% from their September peak, almost returning to their July lows.
The bank remains positive on the sector, citing an average estimated free cash flow yield of 29% for 2027-2030 across its coverage universe, excluding China’s ChangXin Memory Technologies (CXMT).
UBS named Samsung Electronics (KS:005930) as its top pick and maintained buy ratings on all stocks in its memory coverage.
Analysts said growing demand for conventional server memory and solid-state drives was helping offset uncertainty surrounding high-bandwidth memory (HBM), used in AI accelerators.
Orders for server DDR5 memory are increasing into 2027, supported by the deployment of AI inference systems, while demand for Nvidia’s (NASDAQ:NVDA) system-on-chip memory modules, known as SOCAMM, has strengthened, analysts stated.
UBS estimates HBM shipments will grow 49% year-on-year in 2027, partly reflecting a shift in memory configurations as manufacturers accommodate competing demand for other memory products.
Concerns about Chinese rival CXMT’s potential adoption of a more advanced DRAM manufacturing technology also appear premature, UBS analysts said. They expect its near-term development to involve research or sample production rather than a major increase in industry supply.
However, UBS flagged the possibility of faster capacity expansion by Samsung, which could add more than 150,000 wafer starts per month of DRAM capacity in 2028. Analysts said this was unlikely to materially affect industry supply and demand before late 2028.
UBS said the recent share-price correction should encourage memory makers to step up shareholder returns, potentially providing further support for valuations.
Original Article
Memory stocks: Is the wall of worry justified?
DeepSeek releases AI model update, opens search APIs to all users
Citi sees resilient demand for Chinese AI IPOs despite valuation concerns
