Securities industry analysis suggests that China’s semiconductor localization policy may actually work in favor of Samsung Electronics (005930.KS). The logic: Chinese-made AI accelerators have lower computational efficiency compared with Nvidia GPUs, meaning more accelerators and memory are required to process the same computational workload.<\/p>\n
KB Securities named Samsung Electronics as the top beneficiary of China’s expanding AI data center investment in a report released on the 28th. Kim Dong-won, head of research at KB Securities, said: “China’s semiconductor localization policy is expected to act as a direct factor creating additional demand for Samsung Electronics memory, rather than reducing it. This is because Chinese-made AI accelerators have lower computational efficiency compared with Nvidia GPUs, requiring more accelerators and memory to handle the same AI computational workload.”<\/p>\n
He particularly assessed that a technology and product competitiveness gap with China’s ChangXin Memory Technologies (CXMT) still exists in the high-performance server DRAM segment. “China’s four major cloud service providers (CSPs) \u2014 Baidu, Alibaba, Tencent, and ByteDance \u2014 are simultaneously requesting not only multi-year long-term supply agreements (LTAs) for memory from Samsung Electronics, but also foundry outsourcing partnerships for production of their self-developed AI accelerators (ASICs),” he added.<\/p>\n
AI data center investment in the United States and China is expected to expand rapidly. KB Securities estimates that cumulative AI data center investment by both countries from 2026 to 2030 will reach $5 trillion (approximately 6,904 trillion won). Over the same period, combined AI data center capacity in both countries is projected to expand 2.1-fold from 95 gigawatts (GW) this year to 201GW by 2030.<\/p>\n
By country, U.S. AI data center capacity is expected to grow 2.2-fold from 55GW in 2026 to 121GW by 2030, while China’s is projected to double from 40GW in 2026 to 80GW by 2030. While hyperscalers such as Google, Amazon, and Microsoft are executing large-scale investments in the United States, AI investment by major Chinese CSPs including Baidu, Alibaba, and Tencent is also expanding rapidly.<\/p>\n
“Global AI data center investment is expected to develop into a two-power structure with the United States leading the market and China accelerating growth,” Kim said. “Accordingly, the AI infrastructure investment cycle appears to be only in its beginning stages.”<\/p>\n
Considering the advancement of AI models, expansion of inference demand, and increased adoption of self-developed AI accelerators, the analysis suggests that not only quantitative data center expansion but also increases in per-server computational performance and memory capacity will unfold simultaneously. Combined U.S.-China AI data center investment is estimated to exceed an annual average of 1,800 trillion won (approximately $1.3 trillion).<\/p>\n
Meanwhile, some market analysis suggests China’s semiconductor self-reliance is progressing faster than expected. According to Chinese financial media outlet Caixin on the 26th, Goldman Sachs projected in a recent report that China’s DRAM localization is advancing rapidly, and that the country could supply 40% of its high-bandwidth memory (HBM) demand domestically by 2028.<\/p>\n
Despite this, South Korean securities firms maintain an optimistic outlook on the semiconductor cycle. Park Sang-hyun, an analyst at iM Securities, said regarding the Bank of Korea’s upward revision of next year’s GDP growth forecast by 0.8 percentage points: “This is because the semiconductor boom cycle is expected to continue into next year, simultaneously driving improvements in exports, facility investment, and consumption.” He added that considering earnings forecasts for Samsung Electronics and South Korea’s SK Hynix (000660.KS) for this year and next, the semiconductor cycle next year could be even stronger than this year.<\/p>\n","protected":false},"excerpt":{"rendered":"Securities industry analysis suggests that China’s semiconductor localization policy may actually work in favor of Samsung Electronics (005930.KS).…\n","protected":false},"author":2,"featured_media":134796,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18],"tags":[4221,39969,48705,44263,44914,466,15606,2295,127,276,40609],"class_list":["post-134795","post","type-post","status-publish","format-standard","has-post-thumbnail","category-samsung-electronics","tag-ai-data-centers","tag-alibaba","tag-baidu","tag-bytedance","tag-changxin-memory-technologies-cxmt","tag-kb-securities","tag-kim-dong-won","tag-nvidia","tag-samsung","tag-samsung-electronics","tag-tencent"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/posts\/134795","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/comments?post=134795"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/posts\/134795\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/media\/134796"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/media?parent=134795"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/categories?post=134795"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/korea\/wp-json\/wp\/v2\/tags?post=134795"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}