Sarawak's semiconductor-grade polysilicon plant will be the first in Southeast Asia able to make the ultra-high-purity silicon that wafer and chip production starts from, according to The Edge Malaysia. IFC, the World Bank Group's private-sector arm, is lending US$125 million to OCI TerraSus for the project in Bintulu, which is designed for 8,000 tonnes a year and targets full commercial operation in January 2029. For buyers in the region it changes nothing about supply this year or next. It changes what a realistic sourcing map looks like at the end of the decade.
What was announced, and who is building it
OCI TerraSus — formerly OCI Malaysia, a subsidiary of South Korea's OCI Holdings — will build the facility through a joint venture with Japan's Tokuyama Corporation, OCI Tokuyama Semiconductor Materials. The plant sits on a 13.4-hectare site at Samalaju Industrial Park, next to OCI TerraSus's existing solar-grade polysilicon operation.
That adjacency is the most important detail in the announcement. The operator already runs the chlorosilane chemistry, utilities and trained workforce that polysilicon production needs, in the same industrial park. This is a purity upgrade built on a working base, not a greenfield entry into one of the hardest materials businesses there is. IFC's participation also frames the project as regional supply-chain infrastructure powered by clean energy, rather than a purely commercial capacity add.
Why semiconductor-grade polysilicon plant capacity is so rare
Polysilicon for solar cells and for semiconductors usually starts from the same route, the Siemens process: metallurgical silicon is converted to trichlorosilane, distilled, then deposited back onto heated silicon rods. The difference is how far purification goes. Solar-grade material is typically around 6N (99.9999%). Semiconductor-grade material is 9N to 11N, where metallic impurities are counted in parts per trillion. Getting there takes more distillation, contamination control across the entire plant, and substantially more electricity per kilogram.
Two consequences follow. First, only a small number of producers — in the US, Germany, Japan, South Korea and, increasingly, China — make semiconductor-grade material at scale, which is why Southeast Asia has so far imported all of it. Second, new capacity goes where power is cheap, stable and increasingly low-carbon. Sarawak's large hydropower base is why polysilicon producers came to Bintulu in the first place, and it is the same reason the upgrade to semiconductor grade is happening there.
What it means for Southeast Asian manufacturers, segment by segment
- Wafer and device makers: the direct beneficiaries in principle, but only after qualification. New polysilicon has to pass customer-specific purity and consistency audits, and that commonly runs well beyond a year after first output. Treat 2029 as the start of qualification lots, not of volume supply.
- OSAT and back-end operations in Penang, Batam and Vietnam's industrial clusters: no direct material change, since they receive processed wafers rather than polysilicon. Over time, though, an upstream anchor in the region tends to pull more of the supporting supply base closer, from specialty gases to high-purity process chemicals.
- Solar module makers: largely unaffected. Sarawak's solar-grade output is already part of their supply picture.
- Buyers of high-purity process chemicals: as semiconductor investment deepens in the region, expect specification, packaging and documentation requirements to tighten for the same purity-controlled materials you already buy.
Scenario view: what to assume, and when
Announced dates for first-of-kind materials plants are targets, not delivery dates. The useful way to read this one is as a sequence of gates — construction, commissioning, commercial start, then qualification with each individual customer — where every gate can slip on its own and only the last one puts material into a wafer supply chain. The table below plans against each gate rather than against the headline date.
| Period | Expected plant status | Practical assumption for SEA buyers |
|---|---|---|
| 2026–2028 | Construction and commissioning | No change to current polysilicon or wafer supply routes; existing import dependence stands |
| 2029 | Targeted start of full commercial operation, 8,000 t/yr design capacity | First material enters customer qualification; volume open to new buyers likely limited |
| 2030 onward | Qualified output, if the schedule holds | A regional origin becomes a realistic second source for some wafer supply chains |
| If delayed | Start-up or qualification slips, a common pattern for first-of-kind plants | Current multi-origin strategy stays as it is; the project is optionality, not a dependency |
Monitoring checklist for procurement and supply planning
- Track construction and commissioning milestones from OCI TerraSus, Tokuyama and IFC against the January 2029 target.
- Watch for offtake announcements. Whether output is committed to existing wafer makers outside the region or offered to regional buyers decides who actually benefits.
- Map your own exposure: which of your inputs trace back to semiconductor-grade silicon, and how many origins stand behind each one.
- Do not rewrite current sourcing on the strength of a 2029 target. Revisit once qualification results are public.
- Review buffer stock and lead-time assumptions for the high-purity materials that remain concentrated in one region today — the new plant does nothing for those.
- If you buy from wafer or device makers rather than supplying them, ask whether their qualification plans include the Sarawak origin. That is the earliest reliable signal that it will reach your own supply chain.
FAQ
When will Sarawak's semiconductor-grade polysilicon plant start production?
Full commercial operation is targeted for January 2029, according to the announcement reported by The Edge Malaysia. Output then has to pass customer qualification before it moves into wafer production at volume.
How big is the Sarawak polysilicon plant?
It is designed for 8,000 tonnes of semiconductor-grade polysilicon a year, on a 13.4-hectare site at Samalaju Industrial Park in Bintulu. It adds a new origin in Southeast Asia; it does not remove the need for existing ones.
What is the difference between solar-grade and semiconductor-grade polysilicon?
Purity. Solar-grade is around 6N (99.9999%), enough for photovoltaic cells. Semiconductor-grade is 9N to 11N, which wafers for integrated circuits require. The extra purification takes more energy and far tighter contamination control, which is why few plants make it.
How DIC supports this
DIC does not supply polysilicon. The problem this project addresses — critical materials concentrated in a few origins — is the one our supply model is built around for manufacturers who depend on purity-sensitive inputs. We source across multiple origins, run VMI programs so agreed buffer stock is held locally rather than on a vessel, and keep sensitive grades in an isolated pharma-grade warehouse where contamination control is part of how material is handled. If regional investment like this is changing how you plan high-purity material supply, our team can review your lead times and origin exposure with you.
Talk to DIC about supply planning →
