報告_GS_台灣科技半導體CorporateDay2026_Day2_20260602

PDF 原檔：260602_gs_TW-tech-semi-day-2_original.pdf

PDF 原檔：260602_gs_TW-tech-semi-day-2_original.pdf

原始內容

Taiwan Technology: Semiconductors: GS Taiwan Computex & Corporate Day 2026 - Day 2 Key takeaways

We hosted our Goldman Sachs Taiwan Computex & Corporate Day during 1-2 June 2026. During day 2 within our Taiwan Semiconductor coverage, we hosted management teams from ASE (3711.TW; Buy), GlobalWafers (6488.TWO; Neutral), KYEC (2449.TW; Buy), and Hon Precision (7769.TW; Buy). We provide our key takeaways from the sessions below.

ASE (3711.TW; Buy)

Investors' questions have been focusing on ASE's leading edge business. On its full-process business, management sees supply as insu ffi cient, and we expect ASE to expand fairly aggressively given the large opportunity not limited to AI into 2027E. For testing, management expect wafer probing to be the primary growth driver for this year, driven mainly by strong demand from foundry outsourcing. Besides wafer probing, ASE also sees Final Testing ramping up in 2H26, adding to the overall leading-edge testing momentum. Overall, to recap, management has previously guided LEAP revenue to reach US$3.5bn in 2026 and sees even stronger incremental revenue growth for the LEAP business in 2027, which implies at least US$1.9bn of LEAP revenue growth, representing c.54% of YoY increase.

For ASE's mainstream business, management sees steady momentum, with Taiwan wire bond capacity being generally full and power-related demand being strong. On pricing, management described the overall environment as favorable and sees no ASP pressure, while expecting average prices to rise over the longer term as chip complexity increases (more I/O, more RDL). In the long-run, we continue to believe ASE's overall GM will continue to improve, supported by growth in leading-edge revenue contribution, more favorable pricing environment, and higher utilization.

GlobalWafers (6488.TWO; Neutral)

Management attributed weaker near-term GM to capacity expansion, with higher depreciation and ramp-up costs as six locations expand simultaneously, but maintained that the expansion is the right decision given strong AI-driven advanced-node demand. Overall, management is seeing demand recovering with AI-driven memory demand, rising auto silicon content per vehicle, and advanced packaging driving compound-semiconductor demand. Management sees 12-inch wafer at full utilization and mature node back at a healthy level.

GWC also highlighted substantial government support, with subsidy receiving from both Italy and US government. On US expansion, management noted that GWC is the

Goldman Sachs does and seeks to do business with companies covered in its research reports. As a result, investors should be aware that the fi rm may have a con fl ict of interest that could a ff ect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. For Reg AC certi fi cation and other important disclosures, see the Disclosure Appendix, or go to www.gs.com/research/hedge.html. Analysts employed by non-US a ffi liates are not registered/quali fi ed as research analysts with FINRA in the U.S.

Bruce Lu

+886(2)2730-4185 | bruce.lu@gs.com Goldman Sachs (Asia) L.L.C., Taipei Branch

Evelyn Yu

+886(2)2730-4187 | evelyn.yu@gs.com Goldman Sachs (Asia) L.L.C., Taipei Branch

Ryan Huang, CFA

+886(2)2730-4084 | ryan.huang@gs.com Goldman Sachs (Asia) L.L.C., Taipei Branch e92c7a75ab8b4efbba794e6b187208c8

only wafer vendor that is expanding into the US and is seeing very strong customer engagement with clients looking to negotiate 8-10 year LTAs. Post expansion, management targets 12-inch wafer to account for around two-thirds of revenue (about half would be advanced node) with 8-inch and below and also compound semiconductor making up the remaining one-third.

Management also framed several longer-term opportunities tied to AI and advanced packaging. In SOI, GWC sees silicon photonics as the main driver, with US expansion focused on 300mm SOI, which will expand the current SOI capacity by 3x. GWC is also developing 12-inch SiC for advanced packaging heat dissipation. We believe its 12-inch SiC could potentially be adopted in AI-related applications with potential revenue contribution in 2028.

KYEC (2449.TW; Buy)

Management shared that its near-term demand outlook with ASIC demand is continuing to rise and major GPU customer's next-generation platform is on track for production. However, smartphone demand seems to be relatively weaker, impacted by heightened memory prices. Overall, management shared they expect AI revenue mix to rise from around 27-28% in 2025 to around 40% in 2026, with further upside into 2027. On capex, management reiterated guidance for 2026 capex to be NT$50bn, roughly 50% is dedicated for cleanroom construction, 25% for testers and the remainder for burn-in and other equipment.

Management framed a structural mix shift toward the leading edge, where leading edge revenue contribution accounted for over 50% of revenue in 2025 and could potentially exceed 65% in 2026, with further upside into 2027, spanning leading GPU, CPU, and ASICs. KYEC sees strong CPU demand, and we see rising scope for a GPU vendor's CPU to reach mass production in 3Q26 alongside another hyperscaler's CPU. Overall, we maintain that 2027E demand could potentially be stronger vs. 2026, driven by stronger-than-expected CPU and ASIC demand, with rising ASP and UTR both supporting higher GM.

Hon Precision (7769.TW; Buy)

Management has reiterated its 2026 capacity expansion target of more than 40% YoY, and it expects both 2027 and 2028 capacity to grow by around 50% YoY, supported by robust demand across CPU, ASIC, GPU, CPO and EV-related applications. Beyond unit growth, we expect ASP expansion to become an increasingly important driver, supported by AOI attachment, large-package handlers, next-generation ATC systems, and higher cold-plate content. As future AI devices move toward larger package sizes and higher thermal requirements, we expect both content per system and overall equipment value to continue increasing over the next several years.

CPO has been a key area that investors have been focused on. Management indicated that current CPO demand is primarily driven by electrical testing for ASIC switch applications, where Hon Precision's existing platforms are already in volume production. According to our industry channel checks, the fi rst phase of order has been placed in 2H25 and started to ship in 1H26, and we assume a second phase of a similar magnitude in 2H26E. Looking ahead, Hon Precision is developing a new optical-electrical co-test platform incorporating optical alignment modules, which management believes could become the next major CPO opportunity.

e92c7a75ab8b4efbba794e6b187208c8

Investment Thesis, Valuation and Risks

Investment Thesis - ASE (3711.TW, ASX)

ASE is the leading provider of semiconductor manufacturing services in assembly and test. We like ASE's leadership position within the OSAT sector and believe it will outperform peers via continued market share gains despite a volatile market dynamic. Furthermore, we now see a more attractive risk reward given 1) gradual demand recovery as inventory correction cycle comes to an end, 2) structural pro fi tability improvement, and 3) new AI business opportunity as an incremental positive.

While its shares are now trading at the high end of its 10-year trading history (P/E: 5-18x), we believe a valuation premium over through cycle average for ASE is justi fi ed given our positive outlook towards its margin-accretive new AI business. We thus see risk-reward skewed to the upside and have a Buy rating on ASE.

Price Target Risks and Methodology - ASE (3711.TW)

Valuation: We are Buy rated on ASE with a 12m TP of NT$620, which is based on a target P/E of 20x (based on the average forward P/E of ASE's peers) applied to our 2028E EPS, and discounted back to 2027E at 13.6% CoE. Our key assumptions for CoE include: (1) 1.5x beta (sourcing Bloomberg), (2) a 4.25% risk-free rate, and (3) a market risk premium at 6.25% (in line with our GS house view). For its ADR, our 12-month TP is US$43 (based on FX conversion of 31.7 and a 10% ADR premium).

Key downside risks to our views: (1) weaker-than-expected end demand; (2) stronger competition and pricing pressure; (3) poor execution/yields; (4) slower-than-expected progress in advanced packaging; and (5) worse geopolitical/FX movements.

Investment Thesis - GlobalWafers (6488.TWO)

GlobalWafers (GWC) is a global leading raw wafer manufacturer. It is a raw materials supplier for wafer fabs (i.e., IDMs, foundries) and is in the upstream segment within the semi value chain; therefore, we believe there is a time gap for raw wafer players (vs. backend OSAT, IC design, foundry, etc.) to react to any end-demand fl uctuation, which also is partially due to GWC's unique business model (high LTA coverage).

We have a Neutral rating on GWC, as we see a subdued demand recovery with the near-term outlook remaining muted. However, the down-cycle has likely bottomed out for GWC, and we continue to expect revenue and pro fi tability to improve gradually through the coming quarters. We are also optimistic on GWC's 12-inch SiC business. We see a possibility for GWC's 12-inch SiC wafer to be adopted in future generations of AI chips and serve as a key heat dissipator, and we expect 12-inch SiC to become a key growth driver for GWC in the long run, as GWC ramps up its self-crystallization capability. With our target price implying limited upside relative to our coverage, we are Neutral rated on the stock.

Price Target Risks and Methodology - GlobalWafers (6488.TWO)

Valuation: We have a 12m TP of NT$620, which is derived by applying a P/E of 20x (average of through-cycle fwd P/E and upcycle fwd P/E) to its 2027E EPS.

Key risks to our views: 1) stronger-/weaker-than-expected end demand recovery and

e92c7a75ab8b4efbba794e6b187208c8

higher LTA default rate, 2) softening/stronger competition from peers resulting in ASP and pro fi tability change, 3) lower-/higher-than-expected fab expansion costs, 4) higher/lower production yields/e ffi ciency, and 5) favorable/unfavorable FX movements or geopolitical developments.

Investment Thesis - KYEC (2449.TW)

As an industry leader in the IC testing market, we believe KYEC will be a direct bene fi ciary of the AI demand boom especially from 2025E and beyond. KYEC is Nvidia's sole IC testing service provider for the mainstream AI GPUs i.e. H100/A100 chip, and we estimate its current AI exposure to be over 35% of its total revenue in 2025. We expect the multi-year visibility from AI GPU, ASICs, CPU, and AI-related networking demand will support KYEC to grow over 41% CAGR in 2025-2028E. At the same time, rising advanced node contribution, higher utilization rate, and the consigned tester model underpin a structurally stronger margin pro fi le. In our view, the combination of sustained AI demand and expanding test content per chip marks a structural earnings regime shift, supporting our Buy rating on KYEC.

Price Target Risks and Methodology - KYEC (2449.TW)

Valuation: We are Buy-rated on KYEC. Our 12m TP of NT$435 is 85% based on a fundamental valuation with our target P/E multiple of 25x applied to our 2027E EPS, and 15% based on a theoretical M&A value of 31x P/E (a 25% premium over fundamental P/E, inline with the industry average) to our 2027E EPS.

Key risks to our views: 1) AI accelerator shipment volatility, 2) slower than expected advanced node mix ramp, and 3) stronger competition and pricing pressure.

Investment Thesis - Hon Precision (7769.TW)

Hon Precision is a leading Taiwan-based semiconductor testing equipment provider specializing in test handler and Active Thermal Control (ATC) system, which are widely adopted in Final Test (FT) and System-Level Test (SLT). Hon Precision enjoys >90% of market share in the FT handler market for AI/HPC applications, enabling it to fully capture growth in both AI GPU and AI ASICs.

We like Hon Precision as we expect its revenue/earnings to accelerate in coming years driven by 1) growth in AI GPU/ASIC volume, 2) larger chip package size, 3) longer testing time, 4) higher power and thermal requirements, 5) rising adoption of SLT among AI ASICs, and 6) shorter chip upgrade cadence. Hon Precision shares are trading at a lower valuation versus its Taiwan/global peers in terms of forward P/E. We view the company as one of the most attractive names within our TW Advanced Semi Testing coverage, with meaningful re-rating potential as the market begins to re fl ect its strong earnings growth outlook.

Price Target Risks and Methodology - Hon Precision (7769.TW)

Valuation: We are Buy rated on Hon Precision. Our 12-month TP of NT$12,000 is based on a target P/E of 32x (based on close peer correlation analysis) applied to our 2028E EPS, and discounted back to 2027E at 13.6% CoE. Our key assumptions for CoE include: (1) 1.5x beta, (2) a 4.25% risk-free rate, and (3) a market risk premium at 6.25%.

Key downside risks to our views: (1) softer AI/HPC demand, (2) slower adoption of SLT

e92c7a75ab8b4efbba794e6b187208c8

among AI ASICs, and (3) intensifying competition.

e92c7a75ab8b4efbba794e6b187208c8

圖片清單（已驗證 2026-07-02）

回補驗證：僅涵蓋已被 lib 頁嵌入的圖片，非全量驗證。