President Trump signed a Section 232 proclamation on August 6, 2026, imposing a 15% ad valorem tariff on polysilicon derivatives (ingots, wafers, solar cells, modules) and establishing minimum import prices (MIPs) effective December 4, 2026. MIPs are set at $21/kg for raw polysilicon, $100/kg for ingots/wafers, $0.22/watt for solar cells, and $0.38/watt for solar modules. Importers must file documentation demonstrating first-sale prices meet or exceed these floors or face additional duties; the Commerce Department will monitor for stockpiling ahead of Dec 4.
The rationale spans both energy security and semiconductor supply: polysilicon is essential for both solar panel manufacturing and semiconductor production, and the US share of global capacity has fallen from 50% in 2005 to less than 2% in 2024. The proclamation frames the action as defending against decades of foreign dumping, particularly from China, and notes that the US semiconductor industry relies on polysilicon supply chains. The measure also authorizes the Commerce Secretary to establish an incentive program offering tariff relief to companies that build or expand domestic polysilicon, ingot, wafer, or cell manufacturing (construction must begin by January 20, 2029).
For AI hardware builders and data center operators, the downstream impact is layered: solar-grade polysilicon tariffs raise the cost of renewable energy infrastructure needed to power AI clusters, while semiconductor-grade tariffs may increase wafer costs if suppliers pass through the price floor. Industry analysts estimate the tariffs will raise solar module prices by $0.05–$0.11 per watt. Traders have 120 days to frontload inventory before Dec 4, but the proclamation includes anti-stockpiling provisions. The combined effect of this tariff regime plus the optical transceiver ban and tight labor/energy markets creates compounding cost and supply-chain friction for 2026–2027 hyperscaler capex budgets.