Why are customers starting to fund suppliers’ capacity expansion?

  • The way customers secure capacity is changing. Goldman Sachs found during company visits that more substrate, PCB and copper-clad laminate (CCL) companies have secured end-customer support for their 2027–2030 investment plans through long-term contracts, expansion subsidies or advance payments. Such arrangements were previously uncommon, and Goldman views them as a sign that supply constraints will intensify.
  • This makes the earnings outlook better than Goldman previously expected: customers covering part of the equipment investment can reduce suppliers’ depreciation burden, while the gradual recognition of subsidies and similar support can also help profits. Not all companies are willing to accept these arrangements. Goldman believes those that do not sign can retain flexibility in choosing customers and projects and benefit from pricing in a tight market.
  • Most of copper foil maker Co-Tech’s new capacity has already been reserved or supported by customers. Goldman expects its customer subsidies to begin being recognized as early as 2027; the profit opportunity after expansion remains constrained by the pace of yield improvement and the scale of capacity expansion by peers.

How can copper-clad laminate upgrades further raise selling prices?

  • Growth is not driven by sales volume alone. In 2024–2026, shipments of M7, M8 and M9 materials increased each year, driving annual year-on-year increases of 25%–50% in the average selling prices of high-end CCL makers. Goldman believes the next upgrade to M10 will sustain this momentum.
  • EMC’s and TUC’s M10/M11 materials are undergoing customer qualification. Goldman expects EMC to begin shipping M10 as early as the second half of 2027, with a significant ramp-up in 2028, while TUC will begin shipments in 2028. M10 is currently quoted at about 800 US dollars per sheet, compared with the two companies’ overall average selling prices of approximately 63 and 46 US dollars in 2026; the increase in overall selling prices depends on the actual share of higher-grade products.
  • The production difficulty of the alternative PTFE route extends the opportunity for existing materials: supply-chain checks indicate that its PCB yield is only 10%–15%, versus 80%–90% for the M9 solution using Q-glass. Goldman expects PTFE mass production no earlier than after the second half of 2028, and therefore has a more positive view of EMC’s and TUC’s share on the existing route over the next two to three years.

Why are ABF substrate materials a new source of profit?

  • Japanese manufacturers have long held more than 95% of the market for ABF-related materials, but their capacity expansion has been slow. As substrate sizes and layer counts increase and production yields decline, material consumption is rising. End customers have accelerated qualification of Taiwanese and mainland Chinese suppliers since the first half of 2026, opening up room for new entrants to gain share.
  • Goldman expects the ABF substrate CCL market to grow from 800 million–1.2 billion US dollars in 2026 to 4 billion–4.5 billion US dollars in 2028. Its selling price is about 1,500 US dollars per full sheet, far above the 130/180 US dollars for M7/M8. Goldman estimates the product’s current gross margin at more than 50%, above its 2026 forecast of 35.2% for EMC’s overall gross margin.
  • EMC began producing high-end ABF substrate CCL in the second half of 2026, and monthly revenue from its highest-end TPU-related products reached 150 million–200 million New Taiwan dollars. The company plans to increase related monthly capacity from 300 thousand sheets to about 1.5 million sheets by the end of 2027. Goldman forecasts a 20%–30% share in 2028; whether the expansion translates into share still depends on ramping up production.

Will CPO weaken demand for high-end PCB materials?

  • CPO shortens electrical signal paths and could reduce specification requirements for some board materials. Goldman expects it to be adopted first in GPU server switch trays in 2028. Switch boards account for about 14% of its forecast 48 billion US dollar CCL market that year; even if all were downgraded to M4/M7, the decline in the overall market size would be less than 10%.
  • Goldman believes pluggable optical modules will remain mainstream, with their longer electrical signal paths continuing to require high-end PCB/CCL. This also supports demand for mSAP boards used in optical modules: Goldman expects the market to grow at a compound annual rate of 87% in 2026–2028. ZDT plans to expand its mSAP capacity to approximately five times its current level over the next two years, and its related products currently have yields above 90%, providing a production base for meeting demand.