AI has pushed the entire copper-clad laminate (CCL) industry chain all the way upstream, and equipment is starting to 'queue up' first.

The demand for AI computing power continues to surge, and the first to be squeezed is no longer end-terminal chips, but rather the upstream high-frequency, high-speed materials and production line equipment. The boom in AI infrastructure is pushing the pressure of capacity expansion in China's chip industry further up the supply chain, especially in areas like testing, packaging, and high-speed interconnects. The growth rate of demand is faster than expected, and order backlogs have already begun to extend into next year.

This time, the most representative is the Copper Clad Laminate (CCL) industry. This report from Wallstreetcn mentions that the new round of large-scale capacity expansion by CCL manufacturers has directly extended the delivery lead time for key equipment from about eight months to up to two years, with equipment order visibility even scheduled into the first quarter of 2028. From this pace, what the market wants to buy is no longer just materials, but capacity itself; and whether capacity can be realized on time is starting to depend on whether equipment delivery can keep up.

More importantly, this round of expansion is not a single-point action, but an acceleration by both Taiwanese and mainland manufacturers. Those named in the report include EMC, TUC, ITEQ, and Shengyi Technology, all of which are advancing new rounds of high-end capacity expansion. Demand comes from multiple directions including AI servers, 800G switches, electric vehicles, and 5G/6G communications. M8-grade materials have gradually become the standard configuration in high-end PCB design. In other words, CCL hasn't suddenly become hot; it's being pushed forward by the higher bandwidth, higher frequency, and higher heat dissipation requirements of the AI era.

What really puts pressure on the supply chain is that the bottleneck has spread from raw materials to the equipment side. Upstream materials like glass fiber cloth and copper foil were already tight before, and now even production equipment is starting to be in short supply, especially Prepreg equipment, which has become a key link constraining capacity release in the current expansion wave. AMI also confirmed that customer expansion pace is significantly faster than before, with overseas factory construction and investments on both sides of the strait advancing simultaneously; the peak shipment from equipment suppliers is expected to be concentrated in 2027 following customer installation progress.

I prefer to understand this market trend as one thing: AI is not just pushing up material demand, but reordering the entire high-end PCB supply chain. Before, when people talked about CCL, they talked about price, specifications, and inventory; now they talk about lead times, capital expenditure, and whether production lines can lock in equipment early. For manufacturers, this means expansion isn't just a matter of wanting to expand; they must integrate materials, equipment, factory space, and customer production scheduling, otherwise capacity will get stuck halfway.

EMC has completed capacity expansion at its Huangshi, Zhongshan, and Penang factories, and plans to invest over NT$100 billion in expanding production lines in Taiwan, mainland China, and Malaysia simultaneously over the next two years, aiming to increase monthly capacity to 9.45 million sheets by the end of 2027, an expansion of over 50% from current levels. This shows that leading manufacturers are no longer just passively chasing orders, but are actively locking in capacity and seizing the window ahead of time.

My judgment is that the most noteworthy aspect of this CCL expansion round is not how hot the demand is, but that the expansion chain has rapidly transmitted from the customer side to the equipment side. Material price increases are just the appearance; what really determines whether expansion can continue later is whether equipment lead times can be fulfilled on time. In the short term, tight equipment supply will tighten capital expenditure pace; in the medium term, whoever can get equipment early and put capacity into operation faster will be more likely to gain a larger share in the next AI infrastructure cycle.

$Taiwan Semiconductor(TSM.US) $NVIDIA(NVDA.US) $Tesla(TSLA.US)