deprecationObservedPublished: 14h ago

Semiconductor Investing Gets More Selective as HBM, CPU and NAND ‘Pinpoint’ ETFs Flood the Market

Semiconductor exchange-traded funds (ETFs), which have traditionally bundled a broad range of semiconductor-related stocks, are becoming increasingly specialized. Following memory-focused products, ETFs centered on high-bandwidth memory (HBM), NAND flash and central processing units (CPUs) have either been launched or are awaiting release. The trend is positive in that it gives investors more choices. However, questions remain over how much differentiation these products can achieve, as many of them hold broadly similar stocks. Image generated by ChatGPT According to the Korea Exchange on Sept. 1, Hanwha Asset Management received a standard code for its “PLUS Korea HBM Semiconductor” ETF from the exchange the previous day and is currently undergoing a listing review. Compared with its flagship “PLUS Global HBM Semiconductor” ETF, the new product narrows its investment scope to South Korea. Given that SK hynix and Samsung Electronics held 58% and 21% of the global HBM market, respectively, as of the first quarter, the two stocks are expected to form the core of the portfolio. Samsung Asset Management is also undergoing a listing review for its “KODEX U.S. AI Memory TOP2 Plus” ETF. The product is expected to be similar to NH-Amundi Asset Management’s “HANARO U.S. AI Memory Semiconductor TOP4+,” which was listed in May. KB Asset Management, meanwhile, listed its “RISE Global AI NAND Memory Semiconductor” ETF on the same day, allocating about 70% of the portfolio to four major NAND players: Samsung Electronics, SK hynix, SanDisk and Kioxia. Kiwoom Asset Management listed the “KIWOOM U.S. CPU Semiconductor TOP4+” ETF in late July, with holdings including AMD and Intel. Samsung Asset Management followed in early August with its “U.S. CPU Semiconductor TOP10” ETF. In addition, four hybrid semiconductor ETFs have been listed since the second half of the year, combining investments in specific semiconductor companies with bonds. These include the “PLUS SK hynix SanDisk Bond Mix 50” and “1Q Nvidia Alphabet U.S. Treasury Bond Mix 50” ETFs. Image generated by ChatGPT Asset managers are rolling out increasingly specialized semiconductor ETFs as advances in artificial intelligence (AI) expand the semiconductor industry and make it harder to view the sector as a single, homogeneous market. In the past, companies such as Samsung Electronics, SK hynix, Nvidia and TSMC were often grouped together under the broad semiconductor theme. More recently, however, the factors driving share prices and earnings have diverged across segments such as memory, CPUs, GPUs and foundries, increasing the need for a more segmented investment approach. Broadly, the semiconductor industry can be divided into memory products, including DRAM, HBM and NAND, and non-memory segments such as CPUs, GPUs and foundries. The continued inflow of money into semiconductor ETFs is also encouraging the launch of these more “targeted” products. According to the Korea Exchange, semiconductor ETFs accounted for 49.3 trillion won, or 31.9%, of the 154.3 trillion won increase in total ETF net assets this year as of the end of August. The figure covers 92 ETFs whose names include either “semiconductor” or the name of a semiconductor company. Semiconductor ETFs represent only about 8% of the total number of ETF products, but they accounted for roughly one-third of the increase in net assets. For asset managers, this strong investor demand creates an incentive to differentiate new products by narrowing their investment focus according to products, manufacturing processes and different parts of the semiconductor value chain. “In the past, memory was largely viewed as a commodity segment at the lower end of the semiconductor supply chain, but it is now emerging as a strategic asset,” an executive at an asset management firm said. “CPUs, meanwhile, were once considered to be losing relevance after the rise of GPUs, but they are attracting renewed attention with the spread of agentic AI.” “Investors continue to favor semiconductor stocks, while the semiconductor market itself is expanding,” the executive added. “As a result, the ETF market is seeing more attempts to slice semiconductor exposure into different products, processes and parts of the value chain.” Image generated by ChatGPT However, questions remain over whether these increasingly specialized semiconductor ETFs can deliver strong returns amid weakening investor sentiment toward the sector. The Philadelphia Semiconductor Index, often viewed as a bellwether for the global chip industry, has fallen 11% over the past three months. Although it rebounded 2% over the past month, the mood is markedly different from the first half of the year, when the index surged more than 60%. Over the past three months, shares of memory chipmakers have declined sharply, including Samsung Electronics (-25.5%), SK hynix (-28.1%), Micron (-7.4%) and Kioxia (-30.0%). Non-memory semiconductor stocks have also weakened, with Nvidia down 1.6%, TSMC 4.7%, AMD 7.7% and Broadcom 19.5%. There are also doubts over whether the new products can meaningfully differentiate themselves from existing semiconductor ETFs. Despite targeting different themes — NAND and HBM — the “RISE Global AI NAND Memory Semiconductor” and “PLUS Global HBM Semiconductor” ETFs have very similar portfolio compositions. Samsung Electronics, SK hynix, Micron, SanDisk and Lam Research together accounted for 73.6% of the RISE ETF and 89% of the PLUS ETF. While the weightings differ, all five companies rank among the largest holdings in both funds. “Just as important as breaking the semiconductor industry into more detailed segments is deciding which stocks to include,” said Lee Sun-yup, head of AFW Partners. “As different segments now move differently, the portfolio needs to be constructed in a way that properly reflects the theme being targeted.” “Investment options may be limited if managers rely only on Korean stocks,” Lee added. “Since the Korean market is heavily weighted toward memory chipmakers, a strategy that includes global stocks appears necessary to gain exposure to other areas such as CPUs.” sjyoon@chosunbiz.com

Download social card
Copy launch post

Why this byte is shareable

Signal quality

observed

Confidence badge and source context included.

Entity anchor

AI News

Clear company or model context for distribution.

Export ready

1200 x 630 card

Optimized for X, LinkedIn, and chat previews.

Why it matters

Deprecations can break production agents quickly. Teams should audit dependencies and ship migration patches before cutoffs.

Suggested launch post

Use this in X threads, community posts, internal team chats, or launch recaps.

Semiconductor Investing Gets More Selective as HBM, CPU and NAND ‘Pinpoint’ ETFs Flood the Market

Why it matters: Deprecations can break production agents quickly. Teams should audit dependencies and ship migration patches before cutoffs.

Source: It조선
https://a2zai.ai/bytes/...
Post to X
Copy text

Permalink: https://a2zai.ai/bytes/semiconductor-investing-gets-more-selective-as-hbm-cpu-and-nand-pinpoint-etfs-fl-4d9ac2f5

Social card: https://a2zai.ai/bytes/semiconductor-investing-gets-more-selective-as-hbm-cpu-and-nand-pinpoint-etfs-fl-4d9ac2f5/opengraph-image

Social and community

Discussion