Silicon — material dossier

Structural facts + context. Verified components marked.

What it is

Metalloid, atomic number 14. Second most abundant element in the earth's crust — not scarce in nature. The strategic question is never about whether silicon exists in reserves; it's about refining purity. Three product grades matter:

1. Silicon metal (metallurgical, 98-99%) — feedstock for aluminum alloys, silicone chemicals, and further refining. Dominant use by volume. 2. Polysilicon (solar-grade, 6N purity, 99.9999%) — used in photovoltaic cells. Produced via Siemens process or fluidized-bed reactor (FBR). Chinese-dominated. 3. Electronic-grade polysilicon (9N-11N purity, 99.9999999%+) — semiconductor wafers. Tiny volumes but orders-of-magnitude higher price. Dominated by Shin-Etsu, SUMCO, Siltronic, GlobalWafers, Hemlock, Wacker.

Price state (verified 2026-04-24)

Polysilicon is in a bear market. Unlike lithium, neodymium, copper, and cobalt — all rallying on concentrated supply + rising demand — polysilicon has the opposite story: massive Chinese overcapacity meeting softer-than-expected solar installation demand. The 35% three-month decline reflects capacity adds by Tongwei, Xinte, Daqo, GCL-Poly racing each other into oversupply.

Silicon metal (met-grade) also generally weak, tracking polysilicon dynamics and soft aluminum demand.

Where it comes from (verified via Statista)

Silicon metal production (2025)

• China: dominant ("roughly 14× higher than Russia")
• Russia: #2
• Rest: Norway, US, Brazil, France (all small relative to China)
• (Specific tonnages behind Statista paywall; USGS MCS would

have them — future wake with PDF parsing)

Polysilicon production (pre-wake structural knowledge)

• China ~85-90% of global polysilicon capacity
• Top Chinese producers: Tongwei, Daqo New Energy, Xinte

Energy, GCL-Poly, East Hope (ranked by capacity)

• Non-Chinese producers: Wacker (Germany), OCI (Korea +

Malaysia), Hemlock Semiconductor (US — mostly semi-grade, sold solar business)

• US/EU solar-grade production is essentially gone since

the 2011-2014 solar panel trade war; residual capacity serves semi industry

Semi-grade silicon (electronic-grade)

• Polysilicon to wafer: Shin-Etsu, SUMCO, GlobalWafers,

Siltronic, SK Siltron dominate globally

• Taiwan, Japan, Korea = wafer heartland
• Germany (Siltronic), US (GlobalWafers US ops) smaller but

strategic

• Integration with foundry ecosystem (TSMC, Samsung, Intel,

SMIC) is what makes this geography matter

Demand structure

• Silicone chemicals (silanes, silicones, elastomers): ~40%

of silicon-metal demand

• Aluminum alloys: ~30%
• Solar polysilicon: ~25% (most strategically watched)
• Semiconductor electronic-grade: ~5% by volume, much

larger by value

The solar demand headwind

Solar panel prices have collapsed ~70% in 2023-2025 due to Chinese polysilicon/wafer/cell/module oversupply. Several dynamics:

• Solar module deployment continues to grow in volume (GW

installed) but revenue per module has dropped materially

• Chinese installation push has partially absorbed oversupply

but not at margins that would support new capacity

• Installation growth in India, Middle East, LatAm continues
• **Semiconductor demand for silicon is steady but small in

volume**; AI chip demand doesn't move silicon-metal prices (too small a slice)

Policy / geopolitical context

• US IRA / CHIPS Act: solar-panel domestic-content bonuses

push for US/allied polysilicon. Hemlock, REC Silicon (Moses Lake, WA — restarted 2024 partly, Hanwha-owned) benefit.

• EU Solar Pillar of CRMA: attempts to rebuild EU solar

manufacturing (difficult, Chinese cost advantage is structural)

• US anti-dumping / countervailing duties on Chinese

polysilicon (since 2014) — largely ineffective at re-shoring, effective at shifting production to SE Asia

• Xinjiang polysilicon import ban (Uyghur Forced Labor

Prevention Act, 2022) — has meaningfully shifted US import patterns away from Chinese polysilicon

• China: major subsidies for polysilicon capacity

(continue despite oversupply — domestic employment + solar deployment goals)

• EU Carbon Border Adjustment Mechanism (CBAM): silicon

metal + aluminum = carbon-intensive; CBAM could re-shape EU import patterns 2026-2030

Concentration risks

1. Chinese polysilicon oversupply continues (not a shock but a persistent bear drag) — prices could re-test $4/kg lows 2. Xinjiang supply chain uncertainty — ~40% of Chinese polysilicon is from Xinjiang; US import ban already disrupting flows; further western market action possible 3. Semi-grade bottleneck: only ~6 companies globally produce electronic-grade; a production disruption at any would ripple through the semi supply chain (Shin-Etsu Japan earthquake exposure, Hemlock US operational continuity, etc.) 4. Inverse risk: Chinese supply-side discipline. If Beijing forces smaller players to curtail (production cuts, M&A), polysilicon prices could rebound sharply even without demand change

Players to know

Polysilicon (solar-grade)

• Tongwei (China) — world's #1
• Daqo New Energy (China)
• Xinte Energy (China)
• GCL-Poly (China)
• Wacker (Germany)
• OCI (Korea/Malaysia)
• Hemlock Semiconductor (US — semi focus)
• REC Silicon (US — Hanwha-owned, restarted)

Semi-grade wafers

• Shin-Etsu Chemical (Japan)
• SUMCO (Japan)
• GlobalWafers (Taiwan; acquired Siltronic deal blocked

2022, still global #3)

• Siltronic (Germany)
• SK Siltron (Korea)

What to watch monthly

• Weekly polysilicon spot (Bernreuter, PVInsights, SolarMedia)
• China NFMIA monthly supply/demand balance data
• Solar module pricing trends (global + regional)
• US import statistics (is non-Chinese poly share rising?)
• Any Chinese Ministry-led capacity discipline announcements
• Earthquake/disruption news from Japan + Korea semi-wafer hubs
• Q1 earnings from Tongwei, Daqo, Wacker, OCI

Cross-references

• Reports: docs/minerals/reports/YYYY-MM-silicon.md
• Related: connects to country-ETF tech/AI work. Semi-grade

silicon story = Taiwan/Korea/Japan/Germany exposure. See docs/thinking/2026-04-22-tech-ai-value-chain.md.