Why doesn't India produce the ultra-pure silicon its own chip fabs will need?

India is building chip fabs, but no Indian plant makes semiconductor-grade (9N+) polysilicon; domestic projects target only solar-grade material.

Electronic-grade polysilicon production
India's statusNo capability since 2026
Criticalitycritical
Import dependenceover 90% of semiconductor-grade inputs (incl. silicon wafers, specialty gases, high-purity chemicals) imported (2025)
Global makers4
Germany · United States · Japan · South Korea
Typematerials
SectorChemicals & Process
Rests on7 capabilities
Deep-red gaps7
VerificationMachine-checked
Revised2026-07-15

1The gap

India is building its first 300mm chip fab, but not one gram of the ultra-pure silicon those chips are made from is produced anywhere in the country.

Every silicon wafer that becomes a chip starts as polysilicon refined to a purity that few materials on Earth ever reach. Electronic-grade polysilicon is specified at roughly 9N to 10N — around 99.99999999998 percent silicon. Reaching that number is a chemical feat. Metallurgical-grade silicon is first converted into trichlorosilane, a liquid that is then distilled until impurities are stripped to the parts-per-billion level, and finally decomposed inside a Siemens-process reactor so that pure silicon deposits atom by atom onto seed rods. How thoroughly the trichlorosilane is distilled decides everything: the same process yields 7N–8N solar-grade material or the 9N–10N grade a fab will accept.

That difficulty explains why the capability sits with so few. Only four nations — Germany, the United States, Japan and South Korea — host the firms that anchor global supply. The lineage is old and narrow: Wacker, Mitsubishi, Sumitomo Titanium and Hemlock began industrial production between 1959 and 1961, for decades making exclusively electronic-grade material for the semiconductor sector. Today only Tokuyama and Mitsubishi remain focused on electronic grade. When South Korea's OCI shut most of its Gunsan complex in February 2020 under price pressure from low-cost Chinese plants, it kept a single 6,500-tonne line running — precisely because that line makes electronic-grade polysilicon for chips.

India's status is unambiguous: none. Every announced domestic polysilicon project targets solar-grade feedstock, a different and lower-purity product. Adani Group has begun commercial production of 2 GW of solar ingots and wafers in Gujarat and aims to start polysilicon in 2027 at the earliest, reaching 10 GW of metallurgical-grade-silica-to-polysilicon-to-ingot-to-wafer capacity by 2027–28 — all solar grade, not semiconductor grade. Under the 2021 solar-manufacturing PLI scheme, four bidders — Reliance New Energy, Adani Infrastructure, Jindal India Solar and Shirdi Sai Electricals — each proposed a 4 GW plant integrated from polysilicon through wafer, cell and module, again for photovoltaics. In India's polysilicon market, solar PV held a 91.32 percent revenue share in 2025. The upstream fabs will draw on: Tata Electronics, in partnership with Taiwan's PSMC, is building India's first 300mm fab at Dholera for 50,000 wafers a month on 28nm–110nm nodes. Their silicon will be imported.

The gap persists because the capability rests on a stack India has not yet built. Electronic-grade polysilicon needs ultra-low-boron quartz feedstock, trichlorosilane synthesis and distillation, Siemens-process reactors, electronic-grade specialty gases and chemicals, and ppb-level contamination control throughout. Of these, only metallurgical-grade silicon smelting and cleanroom infrastructure are even emerging in India; the rest register as absent. The consequence is measurable: over 90 percent of India's semiconductor-grade inputs — silicon wafers, specialty gases, high-purity chemicals — are imported as of 2025. Global supply concentration compounds the exposure, with China having produced over 90 percent of polysilicon supply in recent years.

2Tech tree

read left to right · click any card for its record

3The builders

Stage = IndiaBUILD assessment from evidence
No builders recorded for this capability yet.

4What it would take

What it would take is the deliberate construction of that upstream stack, not a single plant. India Semiconductor Mission 2.0 explicitly lists indigenous manufacturing of semiconductor chemicals, gases and materials as a goal to reduce import dependence — an acknowledgement that the capability is an aspiration, not an achieved state. India aims for supply-chain independence in critical raw materials only by the early 2030s. The economics are demanding: leading producers have driven Siemens-process cash costs below US$10 per kilogram, the benchmark any new entrant must eventually meet. The distance between solar grade and electronic grade is the distance between a large domestic effort already underway and a capability India has yet to begin.

The Swarajya View
The argument behind this record — in Swarajya.

The diagnosis is free. The argument, the politics, and the case — in Swarajya.

Read on Swarajya →