Global Semis CHIPS IV China Semis self sufficiency builds

EQUITY RESEARCH | August 24, 2026 | 12:06 AM HKT

Global Semis CHIPS IV: China Semis self-sufficiency builds In 2023, amid the challenged constraints on talent, equipment, and advanced-node chips, we set out our first piece (report link) answering the questions of what it would take to build a self- sufficient semiconductor supply chain, how much capital might be required and what are the key factors to drive China’s success in Semis. In this fourth edition after 3 years, our analysis shows:

▪ Ongoing build of supply chain - filling the gaps: Our SPE map shows expansion from Etcher and Deposition to premium models (e.g. high aspect ratio etcher, atomic layer deposition) and more comprehensive categories: Ion Implanter, Inspection, Metrology. Foundry towards 7nm and below with chiplet and advanced packaging to drive production yield rate. AI chips toward 4 PFLOPS in FP8 (4Q28, link) from 320 TFLOPS in FP16 (2023), with SuperPods to drive system-level computing power. Memory towards LPDDR6 (4Q26, link) and HBM3 & 3E (4Q26, report link) from LPDDR4(X) (2024, report link). EDA expands from analog to digital with more tools (e.g. Place and Route; Physical verification) and memory EDA.

▪ Ambitious capex targets that amounts to US$82bn in 2030E, or raised by 79% to capture the growing demand driven by self-sufficiency, generative AI, and customers’ “local for local” strategy. We estimate China 7nm and below logic S/D gap to narrow to 34% by 2035E, and CXMT’s supply to capture 50% of China DRAM demand by 2028E, driven by capacity expansion and yield rate ramp-up.

▪ What China still needs to do: China IC volume self-sufficiency was at 70% in June 2026 (vs. 38% in Jan 2010), while the value S/D gap remains wide given lack of Lithography; R&D and Capex gap vs. global leader remains wide despite strong increases, suggesting more capital required to go global. We introduce China GPU and DRAM TAMs in this report.

▪ Initiating coverage on shares of China’s DRAM leader, CXMT, at Buy with a 12M TP of Rmb129: We expect capacity to more than double by 2030E vs. 2026E to 665k wpm, leading its conventional DRAM supply to reach 41% / 50% of Samsung / SK Hynix by 2028E from 28% / 35% in 2025.

Allen Chang Verena Jeng James Schneider, Ph.D. Shuhei Nakamura Goldman Sachs (Asia) L.L.C. Goldman Sachs (Asia) L.L.C. Goldman Sachs & Co. LLC Goldman Sachs Japan Co., Ltd.

Alexander Duval Giuni Lee Goldman Sachs International Goldman Sachs (Singapore) Pte

Goldman Sachs does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. For Reg AC certification and other important disclosures, see the Disclosure Appendix, or go to Analysts employed by non-US affiliates are not registered/qualified as research analysts with FINRA in the U.S. The Goldman Sachs Group, Inc. Full list of authors inside

Contributing Authors Allen Chang Verena Jeng James Schneider, Ph.D. Goldman Sachs (Asia) L.L.C. Goldman Sachs (Asia) L.L.C. Goldman Sachs & Co. LLC

Shuhei Nakamura Alexander Duval Giuni Lee Goldman Sachs Japan Co., Ltd. Goldman Sachs International Goldman Sachs (Singapore) Pte

Daiki Takayama Yifan Hu Ting Song Goldman Sachs Japan Co., Ltd. Goldman Sachs (Asia) L.L.C. Goldman Sachs (Asia) L.L.C.

Chao Wang Anmol Makkar Kaho Otake Goldman Sachs (Asia) L.L.C., Goldman Sachs & Co. LLC Goldman Sachs Japan Co., Ltd. Taipei Branch

Anant Jakhar Taeyong Lee Al Wang Goldman Sachs India SPL Goldman Sachs (Asia) L.L.C., Goldman Sachs (Asia) L.L.C., Seoul Branch Taipei Branch

CHIPS Act impact Key factors for China to realize self-sufficiency in Semis - leadership, time and capital

CHIPS Act 2 China’s accelerating momentum

CHIPS Act 3 China capex continues to grow; migration ongoing despite challenges