Goldman Sachs - Taiwan Tech: 2026... Goldman Sachs - Taiwan Tech: 2026 PCB/CCL… · @2030FY_X · TwKit

@2030FY_X

Goldman Sachs - Taiwan Tech: 2026... Goldman Sachs - Taiwan Tech: 2026 PCB/CCL Supply Chain Roadshow Core Takeaways - Upgrade Pace Accelerating; High ASP/High Margin ABF Substrate Opportunities Opening Up The Goldman Sachs team conducted in-depth exchanges with 4 companies in the Taiwan PCB/CCL/substrate supply chain and visited over 10 enterprises during the TPCA show in Taipei. Core findings: Long-term terminal demand remains positive, clients are directly locking in 2027-30 new capacity through long-term agreements (LTAs), subsidies, prepayments, and incentive fees, and the structural shortage in the supply chain will be more severe than expected. Consequently, Goldman Sachs raises its mid-to-long-term margin outlook for substrate and PCB/CCL/materials suppliers, and expects further valuation re-rating for related targets. Demand Side: Clients directly step in to lock capacity, supply chain shortage intensifies • Rise of LTA/subsidy model: High-end CCL companies such as EMC, TUC, and Co-Tech have all received requests from terminal clients (cloud providers/chip giants) for long-term agreements, capacity expansion subsidies, prepayments, and incentive fees for exclusive capacity/factories. This phenomenon is appearing for the first time in history. Goldman Sachs believes this confirms the certainty of upward revisions in demand for high-end PCB/CCL/materials over the next few years, and direct funding from clients will bring two major financial benefits: (1) reduced equipment investment leading to a lower depreciation-to-revenue ratio; (2) recognition of prepayments/subsidies directly boosting profits. • Pricing power for non-LTA companies: Companies that do not accept client binding agreements (such as ABF substrate spot market players) will enjoy more flexible client/project selection rights. The long-term pricing environment is actually better, and spot prices possess upside elasticity. • Structural changes: Direct participation of terminal clients in capacity investment will strengthen investor confidence in a multi-year demand/pricing/margin upcycle, benefiting Taiwanese PCB/CCL and materials suppliers that have already had capacity booked by clients. CCL Upgrade Path: M10 entering verification, M11 ready to launch, PTFE progress delayed • Spec iteration accelerating: After M7 volume ramps in 2024, M8 ramp up in 2025, and M9 becomes mainstream in 2026, the upgrade pace has not slowed down. At the TPCA show, both EMC and TUC released M10/M11 solutions (using Q-glass), which are currently undergoing close verification with terminal clients for next-generation AI projects. • Shipment schedule: EMC will begin shipping M10 products (for specific GPU AI racks) as early as 2H27, with massive volume ramp-up in 2028 as more GPU/ASIC projects adopt them; TUC's M10 products will ship in 2028, covering high-end 3.2T Ethernet switches and ASIC projects. • PTFE-based CCL faces obstacles: The production yield of the PTFE solution is only 10-15% (compared to 80-90% for the M9 Q-glass solution), and the mass production timeline has been delayed to beyond 2H28. Therefore, the switch tray PCB upgrade for Nvidia Rubin Ultra GPU servers starting in 2H27 will highly likely adopt M9/M10 level Q-glass/HC (hydrocarbon) materials instead of PTFE. EMC/TUC's market share over the next 2-3 years is expected to beat expectations. • ASP elasticity: The current unit price of M10 level CCL is approximately $800/sheet, whereas the 2026 average ASP for EMC/TUC is only $63/$46. The room for unit price jumps brought by upgrades is significant. New Opportunities: Taiwanese CCL supply chain entering the high ASP/high margin ABF substrate market • Japanese suppliers lagging in capacity investment: Over the past 10+ years, substrate materials such as ABF CCL/T-glass/ultra-thin copper foil have been dominated by Japanese companies (>95% share). However, demand continues to climb (due to larger substrate sizes, increased layer counts, and lower yields, material consumption doubles every 1-2 years), while Japanese capacity expansion remains slow. Terminal clients have accelerated the certification of the Taiwan and mainland China ABF substrate supply chains starting from 1H26. • EMC takes the lead in breakthroughs: EMC entered the ABF substrate CCL market in early 2026, began mass production in 2H26, and current monthly revenue has reached NT$150-200 million (<1% of the company's September monthly revenue). Goldman Sachs expects EMC's ABF substrate CCL capacity to reach 1.5 million sheets/month by the end of 2027, accounting for 50% of global capacity, with market share expected to reach 20-30% in 2028. The TAM for ABF substrate CCL will jump from $0.8-1.2 billion in 2026 to $4.0-4.5 billion in 2028, accounting for 15%/19%/22% of EMC's 2026/27/28 estimated revenue. • Margin explosion: The unit price of ABF substrate CCL is approximately $1,500/sheet (M7/8 is only $130/$180). The current gross margin has reached 50%+, and in the long term is expected to approach the industry average of 70%+ (Goldman Sachs estimates EMC's overall gross margin for 2026/27/28 at 35.2%/40.7%/43.4%). • Co-Tech layout in ultra-thin copper foil: The Japan-dominated ultra-thin copper foil market will also see new entrants. Co-Tech is preparing R&D and new production line designs, with revenue contribution expected as early as 2H28. CPO threat limited, optical module mSAP demand exploding • CPO adoption risk manageable: CPO will be applied to GPU server switch trays in 2028 at the earliest. Even under the extreme assumption that all switch boards downgrade to M4/M7, the impact on the 2028 CCL TAM would be less than 10%. The 30-layer HDI PCB switch board showcased by ZDT proves that pluggable optical modules remain mainstream; their longer electrical paths still require high-end PCB/CCL support. • Optical module mSAP high growth: ZDT's 800G/1.6T/NPO/XPO optical module mSAP solutions are leading, with capacity expected to expand 5x within two years, and current yields exceeding 90%. The optical module mSAP market 2026-28 CAGR is expected to reach 87%. The gross margin is currently 40-45%, and is expected to break through 50% after price hikes and yield improvements, significantly boosting ZDT's profitability (Goldman Sachs estimates ZDT's 2026/27/28 gross margin at 24.7%/27.6%/31.4%). Key Stock Investment Theses 1. EMC (2383 TW, Buy) • Core logic: High-end CCL leader, dual-driven by M10/11 upgrades and ABF substrate CCL. Clients locking in capacity via LTAs eases depreciation pressure, and margins are entering an upward channel. • Valuation: 12-month price target of NT$10,200, based on 27x 2H27-1H28 expected P/E (consistent with the peak multiple of AI component suppliers over the past 3 years). Current stock price is NT$5,070, implying a potential upside of 101%. 2. TUC (6274 TWO, Buy) • Core logic: Key supplier of M7+ high-speed CCL, M10 products shipping in 2028, capacity to expand by more than 2.5x by 2030, and new capacity has been fully booked by clients. • Valuation: 12-month price target of NT$3,010, based on 22x 2027 expected P/E (+2 standard deviations above the past 3-year industry average). Current stock price is NT$1,420, implying a potential upside of 112%. 3. Co-Tech (8358 TWO, Buy) • Core logic: Global general server copper foil leader (~60% share in 2025), actively expanding HVLP3+ capacity (Malaysia Plant 4), effective capacity will reach 10,000/20,000/30,000 tons by the end of 2027/28/29. Client prepayments lock in capacity, subsidies to be recognized from 2027, and margins continuously improving. • Valuation: 12-month price target of NT$940, based on 22x 2028 expected P/E (+1 standard deviation above the past 10-year upcycle valuation). Current stock price is NT$438.5, implying a potential upside of 114%. 4. ZDT (4958 TW, Buy) • Core logic: Global #1 in PCB market share, the only company mass-producing the full category of ABF/BT/HDI/HLC/mSAP. The China ABF substrate market 2025-27 CAGR reaches 53%, and ZDT's 2027 share is expected to exceed 40%, making it the biggest beneficiary. Foldable phone FPCs and optical module mSAP provide additional increments. • Valuation: 12-month price target of NT$1,025, based on SOTP valuation (ABF and BT substrates at 3.5x 2028 expected P/B; high-growth AI server/switch PCBs at 20x 2028 expected P/E; low-growth consumer electronics PCBs at 8x 2028 expected P/E). Current stock price is NT$567, implying a potential upside of 81%. Key Downside Risks 1. Tech substitution: RCC replacing high-end smartphone HDI designs, early CPO popularization; 2. Trade friction: Tariffs and geopolitical conflicts causing smartphone/server/switch shipments to miss expectations; 3. Intensifying competition: Mainland China peers expanding capacity faster than expected, squeezing pricing power; 4. Yield ramp-up: New capacity yield improving slower than expected, dragging margin realization.

曝光 161 · 评论 0 · 点赞 0 · 书签 1