Nanya Technology (2408) held an earnings conference on the 12th and announced its self-prepared consolidated financial results for the third quarter of 2026. Benefiting from capacity being squeezed by major international memory manufacturers and a sharp price surge driven by a shortage of DDR4 memory, the company reported consolidated revenue of NT$133.649 billion for the third quarter. Its quarterly gross margin climbed to 85.7%, net income after tax was NT$87.627 billion, and quarterly earnings per share (EPS) reached NT$25.40—all record highs. Meanwhile, revenue related to AI infrastructure officially exceeded 20% of the company’s total, and it announced plans for large-scale capital expenditure. From 2026 to 2029, the company expects to invest NT$346.6 billion in the first phase of its new 5A fab, advancing its advanced-process and 3D wafer packaging and testing operations at the Pingtung Science Park.
Nanya Technology’s Third-Quarter Results Set Records for Revenue and Profit
Nanya Technology reported third-quarter consolidated revenue of NT$133.649 billion, up 61.9% from the previous quarter. The increase was primarily driven by a significant rise in the average selling price (ASP) of DRAM, boosting profitability across all three key metrics. Quarterly gross profit was NT$114.482 billion, with a gross margin of 85.7%; operating income reached NT$106.894 billion, with an operating margin of 80.0%. Net income after tax for the quarter was NT$87.627 billion, and earnings per share (EPS) were NT$25.40. Net income after tax for the first three quarters totaled NT$163.877 billion, with EPS of NT$49.28, demonstrating strong momentum in the company’s financial position and operations.
AI Infrastructure Revenue Exceeds 20% as Capacity Expansion Plans Advance
In response to significant shifts in the market, Nanya Technology is actively adjusting its product mix to meet demand related to artificial intelligence (AI). Revenue from AI infrastructure, including servers, networking, storage, and system management, has now officially exceeded 20% of the company’s total revenue. The company is also continuing to help customers develop customized AI products and Wafer-on-Wafer products. To meet long-term market demand, the company announced a capital expenditure ceiling of NT$69.7 billion for 2026. The first phase of the 5A fab is planned to have monthly capacity of 39,000 wafers. Total capital expenditure from 2026 to 2029 is expected to reach NT$346.6 billion, and equipment installation at the new fab is scheduled to begin in the first quarter of 2027.
Advancing Process Technology and Planning Advanced Packaging and Testing Facilities in Pingtung
In technology development and capacity planning, Nanya Technology continues to advance process technologies across generations as well as advanced packaging and testing. Customer validation for the third-generation 10-nanometer-class process (1C) is expected to begin in the first quarter of 2027. Development of the fourth-generation (1D) and fifth-generation (1E) processes, as well as extreme ultraviolet (EUV) lithography, is also proceeding as planned. In addition, the company plans to build a 3D wafer packaging and testing facility at the Pingtung Science Park, with advanced 3D integrated circuit (3D IC) technologies such as through-silicon vias (TSV) and wafer stacking expected to be introduced by the end of 2028. Its product lineup continues to cover a range of specifications, including DDR5 and LPDDR5/5X, to strengthen its competitive position in the memory market.
This article, “DRAM Shortage Drives Nanya Technology to NT$87.6 Billion in Quarterly Profit; AI Revenue Share Officially Exceeds 20%,” first appeared on .
【Further Analysis】
【My Additional Perspective】
Nanya Technology’s impressive surge in profit and jump in gross margin essentially reflect a severe structural shortage of traditional DDR4 and general-purpose DRAM, triggered by the large-scale shift of production lines by the world’s three major manufacturers (Samsung, SK hynix, and Micron) to high-bandwidth memory (HBM). The AI boom is not only driving demand for computing power at its core; its spillover effects are also reshaping supply and demand across the general-purpose memory market. The company’s active investment in TSV and advanced 3D IC processes suggests it is seeking to transform from a cyclical commodity memory producer into a high-value-added supplier of customized AI solutions.
【Market Impact】
- **Taiwan Stock Market**: Nanya Technology (2408) may see a significant valuation re-rating, with the near-term outlook **bullish**. This could also lift memory manufacturers and module makers such as Winbond and ADATA. In addition, the launch of a large-scale capital expenditure program of nearly NT$350 billion provides **bullish** medium- to long-term support for cleanroom and semiconductor equipment stocks such as Marketech International and United Integrated Services.
- **U.S. Stock Market**: Memory giants such as Micron (MU) are also benefiting from a sharp rise in ASPs (average selling prices), making the overall memory chip sector **bullish**. The growing share of revenue from AI infrastructure also indirectly confirms that demand across the computing supply chain, including NVIDIA (NVDA), remains solid.
- **Cryptocurrency Market**: The impact is **neutral**. Soaring memory costs could slightly raise the cost of building decentralized storage systems (such as FIL and AR) or node servers, but are not enough to change the broader trend in which cryptocurrency market liquidity remains the dominant factor.
【What to Watch Next】
A key factor in determining whether the supply-demand gap will persist is whether international manufacturers reallocate production lines to general-purpose DRAM after 2026. Investors should also closely monitor progress on 1C process validation in 2027 and the efficiency of capital deployment for equipment installation at the advanced facility in Pingtung.
【Risk Reminder】
High capital expenditure will bring substantial depreciation pressure in the future. If demand for consumer PCs and smartphones fails to recover, or if slower monetization of AI applications leads cloud service providers to cut capital spending, the memory industry could once again face a severe downturn.