The uptrend in high-end CCL profitability restarts; Buy ratings maintained on EMC and TUC with higher target prices

Overview

This report focuses on Taiwan's CCL industry, especially the two high-end CCL makers EMC and TUC. Goldman Sachs believes specification upgrades in AI servers, switches, and general servers will significantly increase CCL value per system, while high-end capacity expansion will be insufficient to match demand growth, leaving the industry in undersupply through 2026-2028E. As a result, higher high-end CCL prices and improved product mix should continue to drive upward revisions to EMC and TUC gross margin, operating margin, and EPS.

Core views

The core view is that high-end CCL price increases are not merely cost pass-through, but are jointly driven by strong AI demand, supply bottlenecks, and customer willingness to procure high-quality CCL. Goldman Sachs expects total CCL TAM to reach US$57.0bn in 2028E, implying a 2025-2028E CAGR of 53%; among this, high-end CCL TAM is expected to grow at 96% and account for 52%/71%/82% of the global CCL market in 2026/2027/2028E, respectively. Low-end CCL price increases in 1H were driven by shortages of raw materials such as E-glass, but as supply eases and consumer electronics demand remains weak, the pace of price increases may slow after 4Q.

Analysis framework

The report combines industry research, AI CCL supply-demand estimates, TAM forecasts by application, product-grade upgrade pathways, revisions to company gross margin and EPS forecasts, and a P/E valuation framework to assess the earnings elasticity and target price changes of EMC and TUC.

Methodology notes

• industry supply-demand analysisCCL supply-demand model

  comparison between high-end CCL demand growth and supply expansion

  By comparing the 2025-2028E CAGR of 96% for high-end CCL demand with supply expansion of about 22%, the report concludes that the industry will remain in undersupply through 2026-2028E.

• earnings forecastingEPS and margin revisions

  upward revisions to revenue, gross margin, operating profit, and net profit forecasts

  Price increases, product mix improvement, and volume ramp-up of high-end AI projects are incorporated into the 2026-2028E forecasts to derive the magnitude of EPS upgrades for EMC and TUC.

• valuation methodP/E target price method

  target price based on forward P/E

  EMC's target price is based on 27x 2H27-1H28E P/E; TUC's target price is based on 22x 2H27-1H28E P/E, reflecting upward revisions to earnings forecasts.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• EMC

  core beneficiary of high-end AI CCL

  Strengths

  It has leading share in high-end HDI materials and SLP materials, and participates in AI server projects for Nvidia, Google, AWS, and others; a higher mix of high-end products is driving gross margin expansion.

  Weaknesses

  Earnings forecasts are sensitive to the ramp-up of AI projects, realization of price increases, and supply bottlenecks in high-end materials.

  Comparison

  Compared with low-end CCL makers, EMC benefits more from AI-driven high-end price increases and product mix upgrades.

  Risks

  AI server demand comes in below expectations, customer bargaining power strengthens, material costs continue to rise, or high-end supply expands faster than expected.

• TUC

  beneficiary of high-end CCL price increases and product mix improvement

  Strengths

  Starting in 2Q26, it stopped supplying some low-end CCL and improved product mix, while also benefiting from price increases in M7+ high-end CCL.

  Weaknesses

  Its 2026 revenue forecast was cut due to capacity constraints, indicating that near-term shipment capability may still limit growth.

  Comparison

  TUC is valued at 22x forward P/E, below the 27x multiple used for EMC, but it similarly benefits from tight supply-demand conditions in high-end CCL.

  Risks

  Capacity bottlenecks, weaker low-end demand, smaller-than-expected price increases, or slower-than-expected product mix improvement.

• high-end CCL

  bottleneck material in the AI infrastructure supply chain

  Strengths

  Demand is driven by upgrades in AI servers, switches, and general servers, and ASP is typically 2x-8x that of mid- to low-end products.

  Weaknesses

  It is highly dependent on AI capital spending and platform upgrade cycles.

  Comparison

  Compared with low-end CCL, high-end CCL price increases are more driven by demand and scarcity rather than simply pass-through of raw material costs.

  Risks

  AI demand softens, customers shift to alternative materials, or supply expansion exceeds expectations.

• low-end CCL

  cost-driven price increase category

  Strengths

  Prices are supported in the near term by E-glass shortages and some capacity shifting toward high-end products.

  Weaknesses

  Demand mainly comes from consumer electronics, with seasonally weak 4Q and 1Q demand and a weaker long-term demand outlook.

  Comparison

  The sustainability of low-end CCL price increases is weaker than that of high-end CCL, and the pace of price increases is expected to slow after 4Q.

  Risks

  Recovery in E-glass supply and continued weak demand for smartphones and PCs.

Key data

• overall CCL TAM2028E US$57.0bn; 2025-2028E CAGR 53%Goldman Sachs expects 2026/2027E CCL TAM of US$20.0bn/US$34.1bn, respectively.
• high-end CCL demand growth2025-2028E CAGR 96%Significantly higher than the roughly 1% CAGR for mid- to low-end CCL demand.
• high-end CCL supply growth2025-2028E CAGR 22%Below demand growth, supporting the undersupply view.
• high-end CCL market share52%/71%/82% in 2026/2027/2028EGoldman Sachs believes high-end products will become the main source of incremental growth in the global CCL market.
• AI server CCL demand2025-2028E CAGR 155%Expected to account for 70% of total CCL demand by 2028, up from 15% in 2025.
• EMC 2Q26/3Q26 gross margin34.1%/36.5%Up about 4.7ppt and 2.4ppt quarter over quarter, respectively.
• TUC 2Q26/3Q26 gross margin29.9%/32.3%Up about 4.8ppt and 2.4ppt quarter over quarter, respectively.
• EMC target priceNT$9,500, previous NT$6,000Buy maintained, based on 27x 2H27-1H28E P/E.
• TUC target priceNT$2,860, previous NT$1,888Buy maintained, based on 22x 2H27-1H28E P/E.

Impact & implications

If Goldman Sachs is correct, high-end CCL may shift from the logic of traditional cyclical products to that of a bottleneck material in AI infrastructure, with price elasticity and margin durability potentially exceeding market expectations. As key suppliers of high-end AI CCL, EMC and TUC are likely to benefit from customer project ramp-up, M7+/M8+ specification upgrades, and product mix optimization driven by the exit of low-end capacity.

Risks

• Demand for AI servers, ASICs, switches, or general servers comes in below expectations, causing high-end CCL demand to miss expectations.
• New high-end CCL capacity comes onstream faster than expected, weakening the basis for undersupply and price increases.
• Customers are unable to continue passing costs on to end customers, hindering CCL makers' ability to raise prices.
• Costs of key materials such as HVLP4 copper foil, Low DK 2 fiberglass, and E-glass continue to rise sharply, squeezing margins.
• Low-end CCL demand continues to weaken due to soft consumer electronics demand, dragging on overall industry shipments.
• EMC or TUC face capacity constraints, yield issues, or customer project timing that falls short of expectations.

What to watch

• Whether the second round of 10%-15% price increases for M7+ CCL in late 3Q26 to early 4Q26 is implemented.
• CCL shipment progress for new AI projects such as AWS Trainium 3, NVDA Vera Rubin, and GOOGL TPU.
• Whether EMC and TUC actual 2Q26 and 3Q26 gross margins and operating margins exceed BBG consensus expectations.
• How quickly E-glass shortages ease and whether low-end CCL prices slow significantly in 4Q26.
• The pace of high-end CCL capacity expansion and progress in customer qualification.
• Whether CCL value per AI server system continues to rise with upgrades from M7+ to M8+.