PDF 原檔:報告_Citi_鴻海2317_20261030_original.pdf
圖片清單(已驗證 2026-11-02)
| 檔名 | size | 分類 | 親眼所見內容 |
|---|---|---|---|
報告_Citi_鴻海2317_20261030_001.png |
536KB | 真資料圖 | 科技日主舞台投影片:四欄分別為 Components L1~L3(SoC Sputtering、BPIC 晶片板)、Enclosure & Mechanical Ownership L5(Metalworking、Plastic Part、Vaper Chamber、HSG Hinge)、SMT & Module L6-7(Ribbon cable、Fingerprint Module、Top Module、Battery Module、Camera、MLB)、System Assy & Validation L10(CVL Scratch-resistant Coating、FATP),屬消費電子產品的垂直整合示意 |
報告_Citi_鴻海2317_20261030_002.png |
605KB | 真資料圖 | 科技日投影片:AI 伺服器版垂直整合,Components/Modules L1-L3(QD、High Speed Cable、Power Busbar、Sockets & Connectors、CPO/CPO ELSFP Transceiver)、Mechanical Modules L5(Cabinet & Enclosure、Chassis、Heatsink Air/Liquid、XPO to CPC Switch)、SMT L6(PCB+PCBA+Power Modules、Fan、Midplane)、Tray L10(Compute Tray、Switch Tray、CPO Switch) |
報告_Citi_鴻海2317_20261030_003.png |
397KB | 真資料圖 | 現場貨櫃式展示實物照:標示 FOXCONN、NVIDIA、NVIDIA DSX POD,貨櫃內可見三組機櫃並排的多排機櫃配置 |
報告_Citi_鴻海2317_20261030_004.png |
735KB | 真資料圖 | NVIDIA 投影片「NVIDIA DSX POD Delivered by Foxconn」:工廠預製整合預測試;7x Vera Rubin NVL72 racks、3 Spectrum-6 SPX racks;25 exaFLOPs of NVFP4 performance;每貨櫃超過 10M 零組件;現場部署由 6 個月縮至 14 天;右側為貨櫃內機櫃剖面示意 |
報告_Citi_鴻海2317_20261030_005.png |
272KB | 文字卡 | Citi 對鴻海 2317.TW 的評等與目標價歷史(上圖)及 Short-Term View/Catalyst Watch 紀錄(下圖);財務/評等類圖表,不嵌 lib |
原始內容
[讀者資訊已遮蔽] citi
30 Oct 2026 13:07:05 ET │ 13 pages
Hon Hai Precision (2317.TW)
Hon Hai Tech Day 2026 - Moving up the AI infrastructure value chain
CITI'S TAKE
At Hon Hai Tech Day 2026, it highlighted its transformation from a traditional EMS provider into an end-to-end AI infrastructure and technology manufacturing platform. Its partnership with NVIDIA has expanded beyond GPU server assembly into complete Vera Rubin rackscale systems, networking, power delivery, cooling and AI factory deployment. The company demonstrated vertical integration spanning L1 - L3 precision components, L5 mechanical assemblies, L6 PCB and power modules, L10 compute and switch trays, L11 rack integration and L12 SuperPOD/data hall deployment. Hon Hai emphasized its global manufacturing footprint and co-development capabilities. We believe the strategic significance lies in expanding Hon Hai's addressable market beyond server assembly toward higher-value system integration, infrastructure engineering and deployment services, potentially increasing content per AI project and strengthening customer relationships.
- ·Hon Hai's partnership with NVIDIA is extending across GPU, CPUs, LPU, DPUs, networking and integrated compute systems. Management highlighted more than ten dedicated manufacturing facilities supporting Vera Rubin production, underscoring the scale of its preparation. A notable engineering breakthrough is the redesign of compute trays to eliminate conventional cables, liquid-cooling hoses and fans, simplifying assembly, improving reliability and enabling greater automation. Hon Hai also showcased its work on robotic assembly, leveraging Omniverse digital twins, computer vision and physical AI models to automate complex operations previously requiring manual precision. These include busbar installation and high-density connector insertion. We believe deeper design-for-manufacturing and design-for-automation capabilities should help Hon Hai accelerate production ramps, improve manufacturing yields and as AI systems become increasingly complex. We also believe this will be the core competence for Hon Hai to secure next-generation rack-scale systems, where manufacturing complexity and system-level validation increasingly differentiate suppliers.
- ·NVIDIA DSX POD, a factory-integrated, pre-validated modular AI infrastructure solution, provided by Hon Hai. Each POD contains seven Vera Rubin NVL72 and three Spectrum-6 SPX racks, and incorporating more than 10 million components. By completing rack integration, cabling, liquid-cooling connections and functional testing before shipment, the solution aims to reduce onsite deployment from approximately six months to as little as 14 days. NVIDIA expects its partner ecosystem to deliver these solutions beginning in early 2027. Hon Hai is also collaborating with
See Appendix A-1 for Analyst Certification, Important Disclosures and Research Analyst Affiliations
Buy
Price (30 Oct 26 13:30)
NT$245.50
Target price NT$400.00
Expected share price return
62.9%
Expected dividend yield
4.2%
Expected total return Market Cap
67.1%
NT$3,455,596M US$105,977M
Laura (Chia Yi) Chen AC
+886-2-8726-9090 laura.cy.chen@citi.com
Michael Hung +886-2-8726-9092 michael.hung@citi.com [讀者資訊已遮蔽] TECO on modular data centers that integrate IT systems, power distribution and cooling within 40-foot containers, with more than 80% of modular assembly completed at the factory. We believe this model could meaningfully expand Hon Hai's revenue opportunity beyond rack assembly into fully integrated infrastructure, while accelerating time to first revenue-generating AI tokens.
- ·Hon Hai highlighted its expanding role in next-generation AI networking, including NVIDIA's Spectrum-6 SN6800 CPO switch, delivering 409.6Tbps aggregate bandwidth through 512 ports at 800Gbps. Hon Hai showcased capabilities spanning CPO optical components, ELSFP modules, high-speed connectivity and complete switch-tray integration, positioning itself to capture additional networking content as optical interconnect adoption accelerates. Next, we expect the scale-up CPO with NVL576 to be announced with VRU next year; Hon Hai will remain the sole supplier, and in our view, its first-mover advantage will be supported for further revenue and earnings upside.
- ·Other than VR200 system, Hon Hai also demonstrated Nvidia's LPUs, integrated into its Vera Rubin ecosystem through LPX inference racks containing 256 LPUs. Unlike GPUs optimized for parallel computation, LPUs accelerate latency-sensitive token generation, particularly feed-forward and mixture-of-experts execution, complementing Rubin GPUs handling prefill and attention workloads. We believe the heterogeneous GPU-LPU computing could expand Hon Hai's addressable market in AI infrastructure integration.
- Beyond AI Data Centers: Hon Hai highlighted progress in EV, autonomous mobility, robotics and smart healthcare. In EVs, it introduced the affordable Model A multi-purpose vehicle, targeting Japan and Southeast Asia. Its CDMS strategy focuses on shortening development cycles and improving cost competitiveness. In robotics, Hon Hai demonstrated physical AI, humanoid robots and digital-twin-enabled manufacturing automation, supported by its collaboration with NVIDIA and ABB Robotics. The company also showcased emerging capabilities in SiC semiconductors, low-Earth-orbit satellites and FoxBrain AI models. We believe these initiatives reinforce Hon Hai's longer-term transformation into a diversified technology manufacturing platform.
Figure 1. Hon Hai's strong vertical integration capability - 1
Source: Company Reports
[讀者資訊已遮蔽]
Figure 2. Hon Hai's strong vertical integration capability - 2

Source: Company Reports
Figure 3. Nvidia's newly announced DSX POD provided by Hon Hai - 1
Source: Company Reports
[讀者資訊已遮蔽] Source: Company Reports [讀者資訊已遮蔽]
Figure 4. Nvidia's newly announced DSX POD provided by Hon Hai - 2

Hon Hai Precision
Valuation
Our target price of NT$400 is based on 18x average 2026E-27E EPS, which is at the higher end of the historical trading range. We believe this multiple is justified by the company's enhanced sales growth and profit margin outlook, as well as new endeavors into AI server, EV, digital health and robotics. Hon Hai has been able to generate positive earnings streams, and thus we believe a P/E valuation methodology is appropriate.
Risks
Downside risks that could impede the stock from reaching our target price include: 1) slower-than-expected iPhone demand; 2) intensified peer competition; 3) weaker-than-expected execution on new business progress; and 4) higher-than-expected impact of COVID-19 or other macro factors.
[讀者資訊已遮蔽] If you are visually impaired and would like to speak to a Citi representative regarding the details of the graphics in this document, please call USA 1-888-500-5008 (TTY: 711), from outside the US +1-210-677-3788