Sumitomo Bakelite plans packaging-material expansion in China and Singapore

Sumitomo Bakelite says MUF fills fine gaps and covers solder joints and the package in a single molding process. It describes compression-molding granules as materials that can fill narrow gaps and support uniform molding, including on large panels. The company says package designs are becoming denser, thinner and larger, adding to the need for these materials.

At Suzhou, the plan includes an additional building and lines at a plant completed in 2024. Production there is scheduled to start in fiscal 2030. In Singapore, the new lines are planned inside the existing plant, with production scheduled to start at the end of fiscal 2028, according to the company announcement.

The capacity projections cover the group, not China alone. Sumitomo Bakelite expects group capacity for MUF materials to rise to about twice its current level and capacity for compression-molding granules to about three times its current level. Those are projections for future capacity, not output already available. The company did not disclose absolute production figures or the expected contribution from each site, and it did not state the expansion’s cost.

The October plan follows a separate Suzhou expansion announced in June. Sumitomo Bakelite said that project would increase its China production capacity by about 30%, with commercial operation expected around December 2028. The company says it began semiconductor encapsulation-material production in China in 1997, added lines in March 2022 and started production at a new plant in 2025. Its June notice does not clarify how that project relates to the October capacity projections, so the 30% figure should not be combined with the group-wide multiples.

How much of the expansion is about AI?

Sumitomo Bakelite attributes the expansion to demand growth for advanced semiconductors, including those used in generative AI and communications satellites. It says this demand is increasing particularly in China and Southeast Asia, while package designs are also becoming more complex in ways that call for additional encapsulation materials.

The company’s June explanation pointed to several other Chinese markets, including internet-of-things devices, 5G applications and electric vehicles. It also cited demand for GPUs, memory and power semiconductors used in AI data centers. Taken together, the company’s statements show that it sees AI as one relevant source of demand, not that AI accounts for a known share of the new capacity.

There is evidence of pressure in the broader packaging market. Semiconductor research firm TrendForce reported in April that AI demand had tightened capacity in advanced packaging, with constraints extending to packaging materials and other components. That industry analysis does not measure demand for Sumitomo Bakelite’s products or establish that AI is driving this specific investment.

The company has not identified customers, package types or production lines that will use the added materials. Nor has it split anticipated demand among AI, communications satellites and other applications. Without those details, the AI connection remains part of management’s explanation for the market opportunity, rather than a quantified account of what will fill the planned capacity.

What could the plans change for chip supply chains?

The immediate potential change is additional planned capacity at two production locations. Sumitomo Bakelite says it already has semiconductor encapsulation-material production sites in Japan, China, Singapore, Taiwan and Belgium. Adding capacity in Suzhou and Singapore could increase the locations from which the company supplies materials, consistent with its stated aim of strengthening supply in major advanced-semiconductor manufacturing regions.

But these planned additions do not by themselves demonstrate broader diversification across the chip supply chain. The announcements do not show that they will reduce shortages or change where chips are made. The materials serve semiconductor packaging; whether added production affects supply depends on how much becomes available and whether it is used by package manufacturers. The company has not disclosed customer orders.

The schedule also limits what the announcement means for near-term supply. Production in Singapore is planned for the end of fiscal 2028, and production in Suzhou for fiscal 2030. Until those lines begin production, the projected capacity increases describe the company’s plans, not materials already available to chipmakers.

A planned acquisition could add another source of production capability, although its contribution is unclear. Kyocera said in a January filing that it had agreed to transfer its chemical business, including semiconductor-related encapsulation materials and pastes, to Sumitomo Bakelite. Kyocera set an end-of-October 2026 target for the transfer, while Sumitomo Bakelite’s October announcement still described the acquisition as planned. The notices do not say how much capacity the business would add or whether it is included in the October projections.

The supply-chain significance will depend on information the company has not yet provided: the eventual output at each site, how the October and June projects fit together, and what share of demand comes from AI-related applications. For now, the announcements establish where Sumitomo Bakelite intends to expand and when production is scheduled—not how much AI-driven demand that capacity will serve.